Papers with Green & Sustainable NLP
CWSeg: An Efficient and General Approach to Chinese Word Segmentation (2023.acl-industry)
Copied to clipboard
| Challenge: | Existing methods for Chinese word segmentation have achieved state-of-the-art performance, but they pose challenges in the deployment. |
| Approach: | They propose to augment PLM-based Chinese word segmentation schemes by developing cohort training and versatile decoding strategies. |
| Outcome: | The proposed model can be used to augment existing PLM-based models and improve their performance on Chinese LLaMA and Alpaca datasets. |
Computationally Efficient Wasserstein Loss for Structured Labels (2021.eacl-srw)
Copied to clipboard
| Challenge: | Existing approaches to estimate the probability distribution of labels are based on tree-Wasserstein distance. |
| Approach: | They propose a tree-Wasserstein distance regularized LDL algorithm for hierarchical text classification tasks. |
| Outcome: | The proposed method performs well on synthetic and real-world datasets and compares favorably with the Sinkhorn algorithm in terms of computation time and memory usage. |
Streamlining Biomedical Research with Specialized LLMs (2025.coling-demos)
Copied to clipboard
| Challenge: | Using large language models, we can generate accurate, context-aware responses with minimal prompts. |
| Approach: | They propose a system that integrates domain-specific large language models with advanced information retrieval techniques to deliver comprehensive and context-aware responses. |
| Outcome: | The proposed system improves quality of dialogue generation and improves efficiency in the biomedical and pharmaceutical domains. |
Multi-News+: Cost-efficient Dataset Cleansing via LLM-based Data Annotation (2024.emnlp-main)
Copied to clipboard
| Challenge: | Various attempts to correct noisy data in the construction process have been made, but human annotation is expensive and time-consuming. |
| Approach: | They propose to use large language models for data annotation to imitate human annotation and classify unrelated documents from a multi-document summarization task. |
| Outcome: | The proposed method imitates human annotation and classifies unrelated documents from the Multi-News dataset. |
MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing (2026.acl-industry)
Copied to clipboard
Junbo Niu, Zheng Liu, Zhuangcheng Gu, Bin Wang, Linke Ouyang, Zhiyuan Zhao, Tao Chu, Tianyao He, Fan Wu, Qintong Zhang, Zhenjiang Jin, Guang Liang, Rui Zhang, Wenzheng Zhang, Yuan Qu, Zhifei Ren, Yuefeng Sun, Zirui Tang, Boyu Niu, Yuanhong Zheng, Dongsheng Ma, Ziyang Miao, Hejun Dong, Siyi Qian, Junyuan Zhang, Fangdong Wang, Jingzhou Chen, Xiaomeng Zhao, Liqun Wei, Wei Li, Shasha Wang, RuiLiang Xu, Yuanyuan Cao, Lu Chen, Qianqian Wu, Huaiyu Gu, Lindong Lu, Dechen Lin, null Shenguanlin, Xuanhe Zhou, Linfeng Zhang, Yuhang Zang, Xiaoyi Dong, Jiaqi Wang, Bo Zhang, Lei Bai, Pei Chu, Weijia Li, Jiang Wu, Lijun Wu, Zhenxiang Li, Guangyu Wang, Zhongying Tu, Chao Xu, Kai Chen, Bowen Zhou, Dahua Lin, Wentao Zhang, Conghui He
| Challenge: | Document images are characterized by higher resolutions, denser content, and more complex structural layouts. |
| Approach: | They propose a 1.2B-parameter document parsing vision-language model that decouples layout analysis from local content recognition. |
| Outcome: | The proposed model surpasses general-purpose and domain-specific models on multiple benchmarks while maintaining significantly lower computational overhead. |
NLP Workbench: Efficient and Extensible Integration of State-of-the-art Text Mining Tools (2023.eacl-demo)
Copied to clipboard
| Challenge: | NLP Workbench is a web-based text mining platform that allows non-expert users to obtain semantic understanding of large-scale corpora using state-of-the-art text mining models. |
| Approach: | They propose to use a microservice architecture to replace existing models or integrate a new one. |
| Outcome: | The proposed model is extensible and can be easily replaced or integrated with existing models. |
High Performance Natural Language Processing (2020.emnlp-tutorials)
Copied to clipboard
| Challenge: | a tutorial on scaling natural language processing will recapitulate the state-of-the-art in the field . |
| Approach: | This cutting-edge tutorial recapitulates the state-of-the-art in natural language processing with scale in perspective. |
| Outcome: | This cutting-edge tutorial recapitulates the state-of-the-art in natural language processing with scale in perspective. |
FT-MDT: Extracting Decision Trees from Medical Texts via a Novel Low-rank Adaptation Method (2025.emnlp-industry)
Copied to clipboard
| Challenge: | Existing methods for extracting medical decision trees rely on manual annotation . PI-LoRA is a low-rank adaptation method for extract medical decision tree from clinical guidelines and textbooks . |
| Approach: | They propose a low-rank adaptation method for automatically extracting medical decision trees from clinical guidelines and textbooks. |
| Outcome: | The proposed method outperforms existing methods for the Text2MDT task while maintaining a lightweight architecture. |
Gatekeeper to save COGS and improve efficiency of Text Prediction (2023.emnlp-industry)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) require massive amounts of computation and storage, such an approach incurs network and high execution cost. |
| Approach: | They propose a model gatekeeper to stop LLM calls that result in incorrect predictions . they show it can save 46.6% of COGS and improve user experience by not showing incorrect predictions. |
| Outcome: | The proposed model gatekeeper saves 46.6% of COGS and improves user experience . it also improves the suggestion rate of the proposed model by 73% . |
Efficient Transformer Knowledge Distillation: A Performance Review (2023.emnlp-industry)
Copied to clipboard
| Challenge: | Pretrained transformer language models have been gaining popularity in the field of natural language processing . however, there is no study into the intersection of these two fields . |
| Approach: | They propose a method to extract knowledge from transformers to produce high-performing efficient attention models with low costs. |
| Outcome: | The proposed model compression method preserves up to 98.6% of original model performance across short-context tasks and up to 95.8% on long-concept Named Entity Recognition tasks while decreasing inference times by up to 57%. |
LAD: LoRA-Adapted Diffusion (2025.emnlp-demos)
Copied to clipboard
| Challenge: | Autoregressive models dominate text generation but suffer from left-to-right decoding constraints that limit efficiency and bidirectional reasoning. |
| Approach: | They propose a framework for non-autoregressive generation that adapts LLaMA models for iterative, bidirectional sequence refinement using LoRA adapters. |
| Outcome: | The proposed framework adapts LLaMA models for iterative, bidirectional sequence refinement using LoRA adapters. |
A Diverse and Effective Retrieval-Based Debt Collection System with Expert Knowledge (2025.naacl-industry)
Copied to clipboard
| Challenge: | Existing debt collection systems lack script diversity, contextual relevance and coherence due to their complexity. |
| Approach: | They propose a script library based on real-world debt collection conversations and a retrieval based response system for contextual relevance. |
| Outcome: | The proposed system improves script diversity and responds to debtor-collector conversations better through knowledge distillation. |
Arxiv Copilot: A Self-Evolving and Efficient LLM System for Personalized Academic Assistance (2024.emnlp-demo)
Copied to clipboard
| Challenge: | Existing solutions for document QA fail to provide personalized and up-to-date information efficiently. |
| Approach: | They propose to deploy a self-evolving, efficient LLM system that can offer personalized research services, maintaining a real-time updated database. |
| Outcome: | The proposed system saves 69.92% of time after efficient deployment. |
InsBank: Evolving Instruction Subset for Ongoing Alignment (2025.findings-emnlp)
Copied to clipboard
Jiayi Shi, Yiwei Li, Shaoxiong Feng, Peiwen Yuan, Xinglin Wang, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Huan Ren, Yao Hu, Kan Li
| Challenge: | Recent studies emphasize that quality and diversity of instruction data are more crucial than quantity, highlighting the need to select diverse, high-quality subsets to reduce training costs. |
| Approach: | They propose to use a continuously updated repository to integrate the latest valuable instruction data with a progressive evolution framework to evolve InsBank over time. |
| Outcome: | The proposed framework outperforms baselines in InsBank evolution and extracts budget-specific subsets. |
Context-Efficient Retrieval with Factual Decomposition (2025.naacl-short)
Copied to clipboard
| Challenge: | Existing models that use dynamically expanding text can be incorporated into large language models. |
| Approach: | They show that pre-processing external corpus into semi-structured "atomic facts" reduces the size of the context and improves inference efficiency. |
| Outcome: | The proposed form of atomic facts improves on question answering tasks when the amount of retrieved text is limited. |
HotelQuEST: Balancing Quality and Efficiency in Agentic Search (2026.eacl-industry)
Copied to clipboard
| Challenge: | Existing benchmarks for agentic search focus primarily on answer quality, overlooking efficiency factors that are critical for real-world deployment. |
| Approach: | They propose a benchmark for hotel search queries that includes 214 hotel query queries that range from simple factual requests to complex queries. |
| Outcome: | The proposed benchmarks show that LLM-based agents achieve higher accuracy than traditional retrievers, but at substantially higher costs due to redundant tool calls and suboptimal routing that fails to match query complexity to model capability. |
Beyond Grid Search: Leveraging Bayesian Optimization for Accelerating RAG Pipeline Optimization (2026.eacl-industry)
Copied to clipboard
| Challenge: | Finding optimal configurations via grid search is computationally prohibitive, limiting real-world scalability. |
| Approach: | They compare BO with grid search to find optimal configurations for RAG pipelines . their framework explores global pipeline and local component-wise optimization . |
| Outcome: | The proposed approach reduces optimization time by up to 84% while maintaining comparable accuracy. |
AdapterDistillation: Non-Destructive Task Composition with Knowledge Distillation (2023.emnlp-industry)
Copied to clipboard
| Challenge: | Recent work on learning from multiple tasks has shown that adding an extra fusion layer to implement knowledge composition is non-scalable for some applications. |
| Approach: | They propose a two-stage knowledge distillation algorithm to extract task specific knowledge by using local data to train a student adapter. |
| Outcome: | Experiments on frequently asked question retrieval in task-oriented dialog systems validate the efficiency of AdapterDistillation. |
ScaleLLM: A Resource-Frugal LLM Serving Framework by Optimizing End-to-End Efficiency (2024.emnlp-industry)
Copied to clipboard
Yuhang Yao, Han Jin, Alay Shah, Shanshan Han, Zijian Hu, Dimitris Stripelis, Yide Ran, Zhaozhuo Xu, Salman Avestimehr, Chaoyang He
| Challenge: | Large language models (LLMs) are widely used in commercial applications . low latency is crucial due to system latency, query concurrency, and computational resources constraints. |
| Approach: | They propose a system that can be resource-efficiently served by addressing bottlenecks beyond LLM inference . they propose 4.3 speed up over vLLM and 1.5 higher throughput . |
| Outcome: | The proposed system outperforms state-of-the-arts with 1.5 higher throughput . it achieves 4.3 speed up with 64 concurrent requests on Mixtral 8x7B . |
Analogical Structure, Minimal Contextual Cues and Contrastive Distractors: Input Design for Sample-Efficient Linguistic Rule Induction (2026.eacl-long)
Copied to clipboard
| Challenge: | Recent systems that use analogical reasoning require extensive knowledge engineering and even transformer-based models show inconsistent results across complexity levels. |
| Approach: | They propose to implement analogical structure, contrastive learning, and minimal contextual cue principles into large language models that train on English verb alternations. |
| Outcome: | The proposed models learn the alternation rules with high F1 on English verb alternations. |
BMInf: An Efficient Toolkit for Big Model Inference and Tuning (2022.acl-demo)
Copied to clipboard
Xu Han, Guoyang Zeng, Weilin Zhao, Zhiyuan Liu, Zhengyan Zhang, Jie Zhou, Jun Zhang, Jia Chao, Maosong Sun
| Challenge: | Recent years, pre-trained language models (PLMs) have achieved promising results on various NLP tasks. |
| Approach: | They propose an open-source toolkit for big model inference and tuning which can support big model tuning at extremely low computation cost. |
| Outcome: | The proposed toolkit can support big model inference and tuning at extremely low computation cost. |
Parameter-efficient Weight Ensembling Facilitates Task-level Knowledge Transfer (2023.acl-short)
Copied to clipboard
| Challenge: | Recent studies show that large pre-trained language models can be adapted to particular tasks in a parameter-efficient manner. |
| Approach: | They propose to use lightweight parameters to transfer them between tasks to obtain similarity between tasks. |
| Outcome: | The proposed methods show an improvement of 5%8% over baselines and could largely facilitate task-level knowledge transfer. |
Prefix Parsing is Just Parsing (2026.acl-short)
Copied to clipboard
| Challenge: | Existing prefix parsers are typically tied to particular parsing algorithms. |
| Approach: | They propose a prefix grammar transformation that reduces prefix parsing to ordinary parsers . they propose enabling prediction of the next token by computing the next-token weight vector . |
| Outcome: | The proposed method reduces prefix parsing to ordinary parsers without modification . the transformed grammar is only a small factor larger than the input . |
Efficient Out-of-Scope Detection in Dialogue Systems via Uncertainty-Driven LLM Routing (2025.acl-industry)
Copied to clipboard
| Challenge: | Out-of-scope (OOS) intent detection is critical in task-oriented dialogue systems . without effective OOS detection, such inputs could lead to incorrect responses, reduced user trust, and eventual system failures. |
| Approach: | They propose a modular framework that combines uncertainty modeling with fine-tuned large language models (LLMs) their method yields state-of-the-art results on key OOS detection benchmarks . |
| Outcome: | The proposed framework yields state-of-the-art results on key OOS detection benchmarks including real-world OOS data. |
TULAP - An Accessible and Sustainable Platform for Turkish Natural Language Processing Resources (2023.eacl-demo)
Copied to clipboard
Susan Uskudarli, Muhammet Şen, Furkan Akkurt, Merve Gürbüz, Onur Gungor, Arzucan Özgür, Tunga Güngör
| Challenge: | a growing interest in the field of natural language processing is resulting in applications solving NLP tasks. |
| Approach: | They propose to create an open-source platform to share Turkish NLP resources . they propose to use the platform to publish open-sourced Turkish Nlp resources based on a research lab's datasets and tools. |
| Outcome: | The proposed platform is easy-to-use and publishes open-source Turkish NLP resources. |
MoFE: Mixture of Frozen Experts Architecture (2025.naacl-industry)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are characterized by their immense size, often consisting of at least one billion parameters. |
| Approach: | They propose a mixture of Frozen Experts architecture that integrates PEFT and MoE to enhance both training efficiency and model scalability. |
| Outcome: | The proposed architecture outperforms other methods while achieving the highest efficiency. |
Online Difficulty Filtering for Reasoning Oriented Reinforcement Learning (2026.eacl-long)
Copied to clipboard
| Challenge: | Recent advances in reinforcement learning with verifiable rewards (RLVR) show that large language models enhance their reasoning abilities when trained with veriable signals. |
| Approach: | They propose a method for a problem-aware filtering system that maximizes learning efficiency by selecting tasks of intermediate difficulty. |
| Outcome: | The proposed model improves when trained with verifiable rewards, but training efficiency is bottleneck . the proposed model achieves +12% gains in less than half the training steps of standard GRPO . |
MKT: A Multi-Stage Knowledge Transfer Framework to Mitigate Catastrophic Forgetting in Multi-Domain Chinese Spelling Correction (2025.emnlp-industry)
Copied to clipboard
Peng Xing, Yinghui Li, Shirong Ma, Xinnian Liang, Haojing Huang, Yangning Li, Shu-Yu Guo, Hai-Tao Zheng, Wenhao Jiang, Ying Shen
| Challenge: | Chinese Spelling Correction (CSC) is a model that detects and corrects spelling errors in given sentences. |
| Approach: | They propose a model-agnostic model with an evolving teacher model and dynamic distillation weights for knowledge transfer in each domain rather than focusing solely on new domain knowledge. |
| Outcome: | The proposed model-agnostic framework is based on an evolving teacher model and dynamic distillation weights for knowledge transfer in each domain, rather than focusing solely on new domain knowledge. |
Efficiency through Auto-Sizing: Notre Dame NLP’s Submission to the WNGT 2019 Efficiency Task (D19-56)
Copied to clipboard
| Challenge: | Notre Dame Natural Language Processing group applied auto-sizing to the Transformer network to reduce the number of parameters in the model. |
| Approach: | They investigated the impact of auto-sizing on the Transformer network by applying a method to inducing sparsity in parameters. |
| Outcome: | The proposed method eliminated more than 25% of the model’s parameters while suffering a decrease of only 1.1 BLEU. |
TensorOpera Router: A Multi-Model Router for Efficient LLM Inference (2024.emnlp-industry)
Copied to clipboard
Dimitris Stripelis, Zhaozhuo Xu, Zijian Hu, Alay Shah, Han Jin, Yuhang Yao, Jipeng Zhang, Tong Zhang, Salman Avestimehr, Chaoyang He
| Challenge: | Large Language Models (LLMs) have demonstrated remarkable performance across a diverse set of domain-specific tasks. |
| Approach: | They propose a non-monolithic LLM querying system that seamlessly integrates various LLM experts into a single query interface and dynamically routes incoming queries to the most high-performant expert based on query’s requirements. |
| Outcome: | The proposed model improves query efficiency by 40% and costs by 30% while maintaining or enhancing model performance by 10%. |
Sparse Adapter Fusion for Continual Learning in NLP (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing methods for learning continuously face challenges such as inefficient parameter reuse across tasks and catastrophic forgetting when tasks are dissimilar. |
| Approach: | They propose a Sparse Adapter Fusion Method which dynamically fuses old and new adapters to address these challenges. |
| Outcome: | The proposed method outperforms state-of-the-art methods while utilizing less than 60% of the parameters. |
Efficient Meta Lifelong-Learning with Limited Memory (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing natural language learning models fail to continuously learn new tasks as they are re-trained throughout their lifetime. |
| Approach: | They propose a meta-lifelong framework that combines three common lifelong learning principles . they propose to store past examples in episodic memory and replay them at training and inference time . |
| Outcome: | The proposed framework achieves state-of-the-art performance using 1% memory size and narrows the gap with multi-task learning. |
Debate-Feedback: A Multi-Agent Framework for Efficient Legal Judgment Prediction (2025.naacl-short)
Copied to clipboard
| Challenge: | Comparative experiments show that our model outperforms several general-purpose and domain-specific legal models. |
| Approach: | They propose a legal judgment prediction model that integrates LLMs with argumentative reasoning techniques to simulate the debate phase of real courtroom trials. |
| Outcome: | The proposed model outperforms several general-purpose and domain-specific legal models and offers a dynamic reasoning process. |
Perceiving the World: Question-guided Reinforcement Learning for Text-based Games (2022.acl-long)
Copied to clipboard
| Challenge: | Text-based games provide an interactive way to study natural language processing. |
| Approach: | They propose a two-phase training framework to decouple language learning from reinforcement learning and improve the sample efficiency. |
| Outcome: | The proposed method significantly improves performance and sample efficiency against compound error and limited pre-training data. |
BIDER: Bridging Knowledge Inconsistency for Efficient Retrieval-Augmented LLMs via Key Supporting Evidence (2024.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) have demonstrated efficacy in knowledge-intensive tasks such as open-domain QA, but inconsistencies between retrieval knowledge and the necessary knowledge for LLMs, leading to a decline in LLM’s answer quality. |
| Approach: | They propose a retrieval-augmented large language model that refines retrieval documents into Key Supporting Evidence (KSE) through knowledge synthesis, supervised fine-tuning, and preference alignment. |
| Outcome: | The proposed model outperforms existing methods in five datasets and boosts answer quality by 7% while reducing input content length by 80%. |
Stop Guessing When to Stop Testing: Efficient Model Evaluation with Just Enough Data (2026.findings-acl)
Copied to clipboard
Ofir Arviv, Kristjan Greenewald, Yotam Perlitz, Hadar Mulian, Michal Shmueli-Scheuer, Leshem Choshen
| Challenge: | Current evaluation practices, typically employing fixed-size benchmarks, are inherently wasteful, continuing to the predetermined sample size even when the CI reaches 2.5, saving 80% of the evaluation cost. |
| Approach: | They propose an adaptive evaluation framework that combines sequential testing with stopping criteria tailored to common evaluation needs such as diminishing returns detection and minimum detectable effect size. |
| Outcome: | The proposed framework reduces computational cost and reliability while maintaining statistical significance. |
Improving Training Efficiency and Reducing Maintenance Costs via Language Specific Model Merging (2026.eacl-industry)
Copied to clipboard
| Challenge: | Recent research on merging multilingual multitask models has shown promise in terms of improved quality, but its computational and maintenance efficiency remains unstudied. |
| Approach: | They propose a method for fine-tuning a multilingual large language model . they compare a "retrain-all" approach and a 'train-once, merge-as-needed' approach . |
| Outcome: | The proposed approach reduces training time by up to 50% while maintaining parity in terms of quality. |
Welcome to the Real World: Efficient, Incremental and Scalable Key Point Analysis (2023.emnlp-industry)
Copied to clipboard
| Challenge: | Key Point Analysis (KPA) extracts the main points from opinions and quantifies their prevalence. |
| Approach: | They propose a key point analysis framework that extracts the main points from opinions and quantifies their prevalence. |
| Outcome: | The proposed system is able to match sentences to key points over five datasets and demonstrate its performance. |
Structural Knowledge Distillation: Tractably Distilling Information for Structured Predictor (2021.acl-long)
Copied to clipboard
Xinyu Wang, Yong Jiang, Zhaohui Yan, Zixia Jia, Nguyen Bach, Tao Wang, Zhongqiang Huang, Fei Huang, Kewei Tu
| Challenge: | Knowledge distillation is a technique to transfer knowledge between models, typically from a large model (the teacher) to a more fine-grained one (the student). |
| Approach: | They propose a factorized form of the knowledge distillation objective for structured prediction which is tractable for many typical choices of the teacher and student models. |
| Outcome: | The proposed model is able to transfer knowledge between teacher and student models without loss of accuracy under four different scenarios. |
Efficient Methods for Natural Language Processing: A Survey (2023.tacl-1)
Copied to clipboard
Marcos Treviso, Ji-Ung Lee, Tianchu Ji, Betty van Aken, Qingqing Cao, Manuel R. Ciosici, Michael Hassid, Kenneth Heafield, Sara Hooker, Colin Raffel, Pedro H. Martins, André F. T. Martins, Jessica Zosa Forde, Peter Milder, Edwin Simpson, Noam Slonim, Jesse Dodge, Emma Strubell, Niranjan Balasubramanian, Leon Derczynski, Iryna Gurevych, Roy Schwartz
| Challenge: | Recent work in natural language processing (NLP) has yielded appealing results from scaling model parameters and training data, but using only scale to improve performance means resource consumption also grows. |
| Approach: | They propose to use data, time, storage, or energy to improve model performance. |
| Outcome: | The proposed methods and findings provide guidance for conducting NLP under limited resources and point towards promising research directions for developing more efficient methods. |
Domain Generalisation of NMT: Fusing Adapters with Leave-One-Domain-Out Training (2022.findings-acl)
Copied to clipboard
| Challenge: | Recent advances in neural machine translation (NMT) research have found NMT models sensitive to distribution shift and adversarial examples. |
| Approach: | They propose a leave-one-domain-out training strategy that learns to combine domain-specific parameters to avoid information leaking. |
| Outcome: | The proposed method outperforms baselines on three language pairs on average. |
Inducer-tuning: Connecting Prefix-tuning and Adapter-tuning (2022.emnlp-main)
Copied to clipboard
| Challenge: | Prefix-tuning is an essential paradigm of parameter-efficient transfer learning . fine-tuned models require separate copies of model parameters for each task . |
| Approach: | They propose to understand and further develop prefix-tuning through the kernel lens . they propose a new variant of prefix tuning that shares the exact mechanism as prefix tun . |
| Outcome: | The proposed method improves prefix-tuning performance by training only a small portion of parameters. |
LLMs on interactive feature collections with implicit dynamic decision strategy (2025.coling-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) struggle to efficiently narrow down the search space . external engineered systems may not fully utilize the inherent problem-solving capabilities of LLMs . |
| Approach: | They propose to implicitly guide Large Language Models to enhance their interactive feature collection abilities within a single prompt. |
| Outcome: | The proposed approach improves the performance of large language models in real-world scenarios. |
EfficientTool: A Cost-Effective Aligning Framework for Tool-Conditioned Agents in SME Scenarios (2026.acl-industry)
Copied to clipboard
| Challenge: | Large language models (LLMs) are increasingly adopted in downstream industries, yet aligning proprietary agents remains challenging. |
| Approach: | They propose a cost-effective, tool-conditioned alignment framework that forms a closed loop over data collection, iterative training, and deployment-oriented evaluation. |
| Outcome: | The proposed framework effectively aligns model in SME scenarios while preserving general tool-calling capability. |
The LLM Already Knows: Estimating LLM-Perceived Question Difficulty via Hidden Representations (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for difficulty estimation rely on repeated response sampling, auxiliary models, or fine-tuning the target model itself. |
| Approach: | They propose a method that leverages only the hidden representations produced by large language models. |
| Outcome: | The proposed method outperforms baselines in difficulty estimation on textual and multimodal tasks and improves adaptive reasoning strategies with fewer generated tokens. |
ComfyUI-Copilot: An Intelligent Assistant for Automated Workflow Development (2025.acl-demo)
Copied to clipboard
Zhenran Xu, Yangxue Yangxue, Yiyu Wang, Qingli Hu, Zijiao Wu, Baotian Hu, Longyue Wang, Weihua Luo, Kaifu Zhang
| Challenge: | ComfyUI-Copilot is a large language model-powered plugin for AI-driven art creation. |
| Approach: | They propose a large language model-powered plugin to enhance the usability of ComfyUI. |
| Outcome: | The new plugin improves the usability and efficiency of ComfyUI . it offers intelligent node and model recommendations and automated one-click workflow construction. |
SWE-Fixer: Training Open-Source LLMs for Effective and Efficient GitHub Issue Resolution (2025.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models excel in code generation benchmarks, but these benchmarks focus on single-file scenarios with constrained context scope. |
| Approach: | They propose an open-source framework to effectively resolve GitHub issues using a code file retrieval module and a model-based code editing module. |
| Outcome: | The proposed approach achieves state-of-the-art performance on two GitHub benchmarks. |
TrendPulse: A Simple yet Efficient Framework for Capturing Viral E-Commerce Spikes via LLM-Driven Contextualization (2026.acl-industry)
Copied to clipboard
| Challenge: | Modern e-commerce platforms mostly depend on reactive discovery, where products surface only after users search for them. |
| Approach: | They propose a framework that identifies regional search momentum and leverages Large Language Model to transform spikes into semantic trends. |
| Outcome: | The proposed framework shows consistent improvements across multiple business metrics and overall user experience. |
Efficient Knowledge Editing via Minimal Precomputation (2025.acl-short)
Copied to clipboard
| Challenge: | Knowledge editing methods like MEMIT require a one-time but significant computational cost. |
| Approach: | They propose to pre-compute 44 million hidden vectors per edited layer . authors show that this precomputation step is unnecessary . |
| Outcome: | The proposed methods can be performed by pre-computing a small portion of 44 million hidden vectors. |
Minimal Ranks, Maximum Confidence: Parameter-efficient Uncertainty Quantification for LoRA (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Low-Rank Adaptation (LoRA) enables parameter-efficient fine-tuning of large language models . standard LoRA lacks mechanisms for uncertainty quantification, leading to overconfident and poorly calibrated models. |
| Approach: | They propose a parameter-efficient Bayesian LoRA method that decomposes weight updates into low-rank matrices. |
| Outcome: | The proposed method achieves strong performance with improved calibration and generalization while maintaining computational efficiency. |
An Efficient Framework for Whole-Page Reranking via Single-Modal Supervision (2026.acl-industry)
Copied to clipboard
Zishuai Zhang, Sihao Yu, null Xiewenyi, Ying Nie, Junfeng Wang, Zhiming Zheng, Dawei Yin, Hainan Zhang
| Challenge: | Existing whole-page reranking methods require large-scale expert annotations to achieve high-quality results. |
| Approach: | They propose a whole-page reranking framework that converts single-modal rankers into page-level guidance by constructing budget-aware candidates for cross-modal annotations and distilling intra-modality preferences to align relevance scales across modalities. |
| Outcome: | The proposed framework reduces annotation costs by 70-90% while outperforming fully-annotated reranking baselines. |
CompAct: Compressed Activations for Memory-Efficient LLM Training (2025.naacl-long)
Copied to clipboard
| Challenge: | Recent studies have focused on reducing peak memory utilization on GPUs, but most work only target the computation graph during training. |
| Approach: | They propose a technique that reduces peak memory utilization on GPUs by 25-30% for pretraining and 50% for fine-tuning of LLMs. |
| Outcome: | The proposed technique reduces peak memory utilization on GPUs by 25-30% for pretraining and 50% for fine-tuning of LLMs. |
Decoding Text Spans for Efficient and Accurate Named-Entity Recognition (2026.acl-industry)
Copied to clipboard
| Challenge: | Named Entity Recognition (NER) is a key component in industrial information extraction pipelines, where systems must satisfy strict latency and throughput constraints in addition to strong accuracy. |
| Approach: | They propose a span-based NER framework that can be used to compute span representations at the final transformer stage, avoiding redundant computation in earlier layers. |
| Outcome: | The proposed framework matches competitive baselines while improving throughput and reducing computational cost. |
Efficient Sequence Learning with Group Recurrent Networks (N18-1)
Copied to clipboard
| Challenge: | Recurrent neural networks have achieved state-of-the-art results in many artificial intelligence tasks, such as language modeling, neural machine translation and speech recognition. |
| Approach: | They propose an efficient architecture to improve the efficiency of such RNN model training by adopting the group strategy for recurrent layers while exploiting the representation rearrangement strategy between layers as well as time steps. |
| Outcome: | The proposed architecture achieves comparable or better accuracy compared with baselines, with a much smaller number of parameters and at a lower computational cost. |
Client-Customized Adaptation for Parameter-Efficient Federated Learning (2023.findings-acl)
Copied to clipboard
| Challenge: | Pre-trained language models have a large memory footprint and are difficult to use in federated learning (FL) |
| Approach: | They propose a hypernetwork-based FL framework that generates client-specific adapters by conditioning the client information. |
| Outcome: | The proposed framework maximizes the utility of shared model parameters while minimizing divergence caused by client heterogeneity. |
Enhancing Efficiency and Exploration in Reinforcement Learning for LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches allocate an equal number of rollouts to all questions during the RL process, which is inefficient. |
| Approach: | They propose a mechanism for dynamically allocating rollout budgets based on the difficulty of the problems, enabling more efficient RL training. |
| Outcome: | The proposed model improves response precision while preserving exploratory ability to uncover potential correct pathways. |
Multilingual Machine Translation with Hyper-Adapters (2022.emnlp-main)
Copied to clipboard
| Challenge: | Multilingual machine translation suffers from negative interference across languages. |
| Approach: | They propose a rescaling fix that reduces the number of parameters and enables training larger hyper-networks. |
| Outcome: | The proposed approach outperforms regular adapters and achieves the same performance with 12 times less parameters. |
An Efficient Context-Dependent Memory Framework for LLM-Centric Agents (2025.naacl-industry)
Copied to clipboard
| Challenge: | a recent study has demonstrated that context-dependent memory encoding can help to retrieve key memory cues essential for problem-solving. |
| Approach: | They propose an efficient architecture miming human memory processes through multistage encoding, context-aware storage, and retrieval strategies for LLM-centric agents. |
| Outcome: | The proposed architecture surpasses state-of-the-art online LLM-centric approaches on two interactive decision-making benchmarks in the navigation and manipulation domain. |
LongSpec: Long-Context Lossless Speculative Decoding with Efficient Drafting and Verification (2026.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) can process extremely long contexts, requiring efficient inference over extended inputs. |
| Approach: | They propose a model that uses a constant-sized key-value cache to train long-context models. |
| Outcome: | Experimental results show that LongSpec achieves 3.26x speedup over strong Flash Attention baselines and 2.34x wall clock time on four math reasoning tasks. |
Towards Efficient LLM Grounding for Embodied Multi-Agent Collaboration (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for grounding large language models suffer from inefficient querying . Existing approaches that rely on physical verification or self-reflection suffer from excessive querying. |
| Approach: | They propose a framework that introduces Reinforced Advantage feedback for efficient self-refinement of plans. |
| Outcome: | The proposed framework surpasses baselines in success rate and significantly decreases interaction steps of agents and query rounds of LLMs. |
JSON Whisperer: Efficient JSON Editing with LLMs (2025.emnlp-industry)
Copied to clipboard
| Challenge: | Large language models can modify JSON documents through natural language commands . current approaches regenerate entire structures for each edit, consuming computational resources . |
| Approach: | They propose a framework that enables large language models to generate diff patches instead of complete documents. |
| Outcome: | The proposed framework reduces token usage by 31% while maintaining edit quality within 5% of full regeneration. |
Confidence-Driven Multi-Scale Model Selection for Cost-Efficient Inference (2026.findings-eacl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have revolutionized inference across diverse natural language tasks, with larger models performing better but at higher computational costs. |
| Approach: | They propose a confidence-driven strategy that dynamically selects the most suitable model based on confidence estimates. |
| Outcome: | The proposed approach reduces token usage by approximately 60% and improves cost efficiency on the Massive Multitask Language Understanding (MMLU) benchmark. |
Thinking with DistilQwen: A Tale of Four Distilled Reasoning and Reward Model Series (2025.emnlp-industry)
Copied to clipboard
| Challenge: | In the rapidly evolving landscape of large language models, the need for efficient reasoning models has become increasingly urgent. |
| Approach: | They extend the Qwen model family by introducing four model series specifically designed for industrial applications. |
| Outcome: | The proposed models outperform previous models in multiple benchmarks and provide scalable training and inference functionality on the Alibaba Cloud PAI platform. |
Dense Template Retrieval for Customer Support (2022.coling-1)
Copied to clipboard
| Challenge: | Templated answers are used to cover a wide range of topics, but the number of templates is often too high for an agent to manually search. |
| Approach: | They propose a dense retrieval framework that adapts a standard in-batch negatives technique to support unpaired sampling of queries and templates. |
| Outcome: | The proposed approach improves performance and training speed over more standard methods. |
Anchor and Broadcast: An Efficient Concept Alignment Approach for Evaluation of Semantic Graphs (2024.lrec-main)
Copied to clipboard
| Challenge: | Abstract Meaning Representation (AMR) is a sentencelevel formalism designed for English. |
| Approach: | They present an intuitive tool for evaluating graph-based meaning representations . they use an anchor broadcast alignment algorithm that is not subject to local maxima . |
| Outcome: | The proposed tool is highly correlated with the widely used Smatch score, but computation takes only about 40% the time. |
Condensing Multilingual Knowledge with Lightweight Language-Specific Modules (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to boost performance in multilingual models but scalability is difficult to manage. |
| Approach: | They propose a method that incorporates language-specific (LS) modules to boost model performance. |
| Outcome: | The proposed method outperforms state-of-the-art methods while outperforming existing methods. |
The Framework Tax: Disparities Between Inference Efficiency in NLP Research and Deployment (2023.emnlp-main)
Copied to clipboard
| Challenge: | Inference is estimated to make up 80 to 90% of ML cloud computing demand . |
| Approach: | They propose to identify bottlenecks in deep learning frameworks that are causing the disparity in model latency as hardware speed increases over time. |
| Outcome: | The proposed models show that the framework tax is increasing as the hardware speed increases over time. |
Towards Efficient and Multifaceted Computer-assisted Pronunciation Training Leveraging Hierarchical Selective State Space Model and Decoupled Cross-entropy Loss (2025.naacl-long)
Copied to clipboard
| Challenge: | APA and MDD are two of the main tasks of computer-assisted pronunciation training (CAPT) systems. |
| Approach: | They propose a computer-assisted pronunciation training approach that integrates APA and MDD tasks in parallel. |
| Outcome: | The proposed approach improves on APA and MDD tasks, and achieves an F1 score of 63.85%. |
Optimal Expert-Attention Allocation in Mixture-of-Experts: A Scalable Law for Dynamic Model Design (2026.acl-industry)
Copied to clipboard
| Challenge: | a novel extension of neural scaling laws to Mixture-of-Experts models is proposed . a ratio of expert-attention compute is crucial for efficient MoE models . |
| Approach: | They propose an extension of neural scaling laws to Mixture-of-Experts (MoE) models . they define the ratio r as the fraction of total FLOPs per token dedicated to expert and attention layers . |
| Outcome: | The proposed model can be tuned beyond size and data with the proposed model. |
Generalized chart constraints for efficient PCFG and TAG parsing (P18-2)
Copied to clipboard
| Challenge: | Existing pruning techniques limit chart constraints to PCFGs and cannot be applied to more expressive grammars. |
| Approach: | They propose to apply chart constraints to more expressive grammars and a neural tagger which predicts chart constraints at very high precision. |
| Outcome: | The proposed technique accelerates both PCFG and TAG parsing by two orders of magnitude while improving accuracy. |
DecEx-RAG: Boosting Agentic Retrieval-Augmented Generation with Decision and Execution Optimization via Process Supervision (2025.emnlp-industry)
Copied to clipboard
Yongqi Leng, Yikun Lei, Xikai Liu, Meizhi Zhong, Bojian Xiong, Yurong Zhang, Yan Gao, null Yiwu, Yao Hu, Deyi Xiong
| Challenge: | Recent advances in outcome-supervised reinforcement learning (RL) have shown strong performance, but this approach still suffers from inefficient exploration, sparse reward signals, and ambiguous global reward feedback. |
| Approach: | They propose a model that models RAG as a Markov Decision Process (MDP) and introduces an efficient pruning strategy to optimize data expansion. |
| Outcome: | The proposed model outperforms existing methods and achieves an average performance improvement of 6.2% across six datasets. |
Lightweight Adapter Tuning for Multilingual Speech Translation (2021.acl-short)
Copied to clipboard
| Challenge: | Adapter tuning is an efficient alternative to fine-tuning in NLP . a multilingual model could be outperformed by its bilingual counterparts . |
| Approach: | They propose to use adapter tuning to optimize for multilingual speech translation . they use pre-trained models to freeze pre-train parameters and inject lightweight modules . |
| Outcome: | The proposed adapters can specialize to specific language pairs with low extra cost . the proposed models outperform bilingual models on high-resource language pairs . |
Adaptive Convolution for Multi-Relational Learning (N19-1)
Copied to clipboard
| Challenge: | Existing convolutional neural networks fail to model full interactions between entities and relations, which limits the performance of link prediction. |
| Approach: | They propose a convolutional network that maximizes entity-relation interactions in a convergent fashion. |
| Outcome: | The proposed convolutional network performs better than baseline models on multiple datasets. |
Scientific Paper Retrieval with LLM-Guided Semantic-Based Ranking (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Recent studies also use large language models (LLMs) for query understanding, but these methods lack grounding in corpus-specific knowledge and may generate unreliable or unfaithful content. |
| Approach: | They propose a paper retrieval framework that combines large language models (LLMs) with a concept-based semantic index to capture scientific concepts. |
| Outcome: | The proposed framework improves the performance of various base retrievers, surpasses strong existing LLM-based baselines, and remains highly efficient. |
Analyzing and Evaluating Correlation Measures in NLG Meta-Evaluation (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing studies have not investigated the differences between different correlation measures in meta-evaluation. |
| Approach: | They analyze 12 common correlation measures using real-world data from six widely-used NLG evaluation datasets and 32 evaluation metrics. |
| Outcome: | The proposed measures exhibit the best performance in discriminative power and ranking consistency . the measures using system-level grouping or Kendall correlation are the least sensitive to score granularity . |
Efficient Agent Evaluation via Diversity-Guided User Simulation (2026.acl-industry)
Copied to clipboard
| Challenge: | Large language models (LLMs) are increasingly deployed as customer-facing agents due to stochastic, multi-turn interactions. |
| Approach: | They propose a snapshot-based, coverage-guided user simulation framework for efficient and systematic exploration of multi-turn agent behavior. |
| Outcome: | Experiments on multi-domain benchmarks show that the proposed method improves failure discovery efficiency and task-level coverage compared to standard linear rollout evaluation without increasing overall cost. |
Reward Difference Optimization For Sample Reweighting In Offline RLHF (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are becoming more capable, but their maximum likelihood objective for the next token prediction falls short in capturing such crucial human values. |
| Approach: | They propose a reward difference prediction method that uses reward difference coefficients to reweigh sample pairs in offline RLHF and a difference model that considers rich interactions between a pair of responses. |
| Outcome: | The proposed method is effective in both automatic metrics and human evaluation. |
Adapter Pruning using Tropical Characterization (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies on adapter pruning have not examined the optimal number of adapter parameters needed for downstream applications. |
| Approach: | They propose an adapter pruning approach that prunes adapter parameters without changing the orientation of underlying tropical hypersurfaces. |
| Outcome: | The proposed approach prunes adapter layers without changing the orientation of underlying tropical hypersurfaces. |
LoRA Meets Dropout under a Unified Framework (2024.findings-acl)
Copied to clipboard
| Challenge: | Parameter-efficientfinetuning (PEFT) has gained popularity as a lightweight approach for model customization. |
| Approach: | They propose a parameter-efficient dropout method that is overfitting-prone and parameter-freezed. |
| Outcome: | The proposed method is superior to existing methods and compares with transformer-specific methods. |
Pruning Adatperfusion with Lottery Ticket Hypothesis (2022.findings-naacl)
Copied to clipboard
| Challenge: | Pre-trained language models are computationally expensive to fine-tune and require large storage. |
| Approach: | They propose a method to identify the influence of each adapter module and a way to prune adapters based on the Lottery Ticket Hypothesis. |
| Outcome: | The proposed model reduces size significantly while keeping performance intact. |
SubRegWeigh: Effective and Efficient Annotation Weighing with Subword Regularization (2025.coling-main)
Copied to clipboard
| Challenge: | Existing methods to reduce the adverse effect of annotation errors are time-consuming because they require many trained models to detect errors. |
| Approach: | They propose a method that uses a tokenization technique called subword regularization to simulate multiple error detection models for detecting errors. |
| Outcome: | The proposed method performs weighting weighting four to five times faster than existing methods and improves in document classification and named entity recognition tasks. |
Unlocking Data-free Low-bit Quantization with Matrix Decomposition for KV Cache Compression (2024.acl-long)
Copied to clipboard
| Challenge: | Existing methods to compress KV cache compromise precision or require extra data for calibration, limiting their practicality in LLM deployment. |
| Approach: | They propose a low-bit quantization technique based on tensor decomposition to effectively compress KV cache. |
| Outcome: | The proposed method reduces memory footprint and performance by 75% . it is compared with existing methods that compromise precision or require extra data for calibration . |
Improving Neural Topic Models using Knowledge Distillation (2020.emnlp-main)
Copied to clipboard
| Challenge: | Current paradigms for transfer learning use general knowledge as a foundation for more specialized endeavors. |
| Approach: | They propose to combine probabilistic topic models and pretrained transformers to improve topic quality by using knowledge distillation. |
| Outcome: | The proposed framework improves topic quality over all estimated topics and in head-to-head comparisons of aligned topics. |
Training Long-Context LLMs Efficiently via Chunk-wise Optimization (2025.findings-acl)
Copied to clipboard
| Challenge: | Recent advances in long-context large language models have demonstrated superior retrieval quality compared to retrievalaugmented generation (RAG) approaches. |
| Approach: | They propose a memory-efficient training paradigm that partitions lengthy inputs into manageable chunks. |
| Outcome: | The proposed model expands maximum sequence length from 1K to 16K tokens on a single RTX 3090 GPU, while SpaCO achieves accelerated training speed. |
Weighted self Distillation for Chinese word segmentation (2022.findings-acl)
Copied to clipboard
| Challenge: | Recent researches show that multi-criteria resources and n-gram features are beneficial to Chinese word segmentation (CWS). |
| Approach: | They propose a framework that uses weighted self distillation to learn Chinese word segmentation using unigram features. |
| Outcome: | The proposed framework achieves state-of-the-art or competitive performance on SIGHAN Bakeoff datasets. |
Leveraging LLMs to Streamline the Review of Public Funding Applications (2025.emnlp-industry)
Copied to clipboard
João DS Marques, Andre Vicente Duarte, André Mendes Marques de Carvalho, Gil Rocha, Bruno Martins, Arlindo L. Oliveira
| Challenge: | Large Language Models (LLMs) have been used to improve evaluation processes, but there are concerns over reliability and potential misapplications. |
| Approach: | They propose to deploy AI-assisted evaluation in two government initiatives . they found that the solution increased reviewer productivity by 20.1% . |
| Outcome: | The proposed solution reduced reviewer productivity by 20.1% while keeping a negligible false-positive rate . the proposed solution reduces evaluation time by more than 2 months . |
A Lightweight Recurrent Network for Sequence Modeling (P19-1)
Copied to clipboard
| Challenge: | Recent studies show that recurrent networks suffer from severe computational inefficiency due to weak parallelization. |
| Approach: | They propose a lightweight recurrent network (LRN) that uses input and forget gates to handle long-range dependencies and gradient vanishing and explosion. |
| Outcome: | The proposed recurrent network yields the best running efficiency on six NLP tasks. |
Mitigating Context Interference for Reliable and Efficient Search Agents (2026.acl-long)
Copied to clipboard
Boyang Xue, Bin Wu, Shuofei Qiao, Sheng Wang, Rui Wang, Yiming Du, Hongru Wang, Jeff Z. Pan, Emine Yilmaz, Kam-Fai Wong, Aldo Lipani
| Challenge: | Recent research empowers Large Language Models (LLMs) as multi-turn search agents to iteratively retrieve and generate outputs until complex tasks are solved. |
| Approach: | They propose a distill-based context refiner to dynamically mitigate context interference . they also propose RLs that refine contexts to generate outputs . |
| Outcome: | The proposed refiner can mitigate context interference in multi-turn search agents. |
SparseAdapter: An Easy Approach for Improving the Parameter-Efficiency of Adapters (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Pretrain-finetuned models are increasingly complex and require more parameters to match the performance of full fine-tuning. |
| Approach: | They propose an efficient Adapter Tuning technique that freezes pretrained language models and fine-tunes a few extra modules. |
| Outcome: | The proposed setting outperforms the standard Adapter Tuning by 80% . the proposed setting is easy to use and has a high sparse ratio . |
FunnelRAG: A Coarse-to-Fine Progressive Retrieval Paradigm for RAG (2025.findings-naacl)
Copied to clipboard
| Challenge: | Retrieval-Augmented Generation (RAG) is widely adopted in Large Language Models, but is flat and has limitations such as a significant burden on one retriever and constant granularity limits the ceiling of retrieval performance. |
| Approach: | They propose a progressive retrieval paradigm with coarse-to-fine granularity for RAG, termed FunnelRAG, so as to balance effectiveness and efficiency. |
| Outcome: | The proposed paradigm achieves comparable retrieval performance while the time overhead is reduced by nearly 40%. |
ADaPT: Token-Level Decoupling for Efficient Large Reasoning Models (2026.findings-acl)
Copied to clipboard
Tingyun li, Zishang Jiang, Jinyi Han, Xinyi Wang, Sihang Jiang, Han Xia, Zhaoqian Dai, Ma Shuguang, Fei Yu, Jiaqing Liang, Yanghua Xiao
| Challenge: | Existing efficiency-oriented methods attempt to shorten or mix reasoning strategies, yet often degrade reasoning capability. |
| Approach: | They propose a token-level dual-process framework that explicitly decouples efficiency and correctness signals during training. |
| Outcome: | The proposed framework reduces inference cost while maintaining strong reasoning ability across multiple benchmarks. |
Operator Selection and Ordering in a Pipeline Approach to Efficiency Optimizations for Transformers (2023.findings-acl)
Copied to clipboard
| Challenge: | Natural language processing tasks rely on complex neural models . transformer-based models are typically slow to execute, making it a non-trivial challenge to apply them in real-world applications. |
| Approach: | They propose to consider an efficiency method as an operator applied on a model . they find that the commutativity and cumulativeness of efficiency operators are plausible . |
| Outcome: | The proposed method is commutative and cumulative, and the results are estimated by combining methods. |
Towards Fast and Accurate Neural Chinese Word Segmentation with Multi-Criteria Learning (2020.coling-main)
Copied to clipboard
| Challenge: | Chinese word segmentation datasets have ambiguous annotation criteria resulting in multi-grained compounds. |
| Approach: | They propose a domain adaptive segmenter to exploit diverse annotation criteria of datasets . they use bidirectional encoder representations from transformers to introduce open-domain knowledge . |
| Outcome: | The proposed model outperforms the state-of-the-art models on 10 Chinese word datasets with superior efficiency. |
RankNAS: Efficient Neural Architecture Search by Pairwise Ranking (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods require training millions of architectures to estimate the accuracy of the search results. |
| Approach: | They propose a performance ranking method (RankNAS) that uses pairwise ranking and search space pruning to enlarge the search space. |
| Outcome: | The proposed method significantly accelerates NAS through pairwise ranking and search space pruning. |
F2TEval: Human-Aligned Multi-Dimensional Evaluation for Figure-to-Text Task (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing evaluation methods for Figure-to-Text tasks are limited due to the inherent ambiguity and semantic compression of figures, the generated texts suffer from factual inaccuracies, incomplete coverage, and weak logical reasoning. |
| Approach: | They propose a five-dimensional reference-free evaluation method aligned with expert criteria to support fine-grained evaluation. |
| Outcome: | The proposed method outperforms Gemini-2.0 and Claude-3.5 with only 0.9B parameters. |
DRP: Distilled Reasoning Pruning with Mathematical Skill-aware Step Decomposition for Efficient Large Reasoning Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing solutions to this problem are inference-time pruning and tuning-based distillation. |
| Approach: | They propose a framework that combines inference-time pruning with tuning-based distillation to enable efficient and accurate reasoning. |
| Outcome: | The proposed framework reduces token usage while improving accuracy on GSM8K and AIME tokens while avoiding performance drop. |
Sparser Mixture-of-Adapters with Cross-Layer Generalization (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing methods for training large language models do not allow sharing adapters across layers . existing methods do not support sharing adapter pools, leading to redundancy and poor generalization . |
| Approach: | They propose a mixture-of-adapter framework that trains a pool of lightweight adapters at each layer and selects the most suitable ones for each input. |
| Outcome: | The proposed framework reduces active adapters by over 85% while boosting task accuracy. |
TokAlign: Efficient Vocabulary Adaptation via Token Alignment (2025.acl-long)
Copied to clipboard
| Challenge: | Tokenization is a foundational step for Large Language Models (LLMs) but low compression rate of vanilla tokenizers decelerates training and inference process. |
| Approach: | They propose a method to replace the vocabulary of Large Language Models (LLMs) by learning a one-to-one mapping matrix for token IDs. |
| Outcome: | The proposed method significantly improves multilingual text compression rates and vocabulary initialization for Large Language Models. |
EcoSafeRAG: Efficient Security through Context Analysis in Retrieval-Augmented Generation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing defense methods rely on internal knowledge of the model, which conflicts with the design concept of Retrieval-Augmented Generation (RAG). |
| Approach: | EcoSafeRAG uses sentence-level processing and bait-guided context diversity detection to identify malicious content . |
| Outcome: | EcoSafeRAG uses sentence-level processing and bait-guided context diversity detection to identify malicious content. |
MBA-RAG: a Bandit Approach for Adaptive Retrieval-Augmented Generation through Question Complexity (2025.coling-main)
Copied to clipboard
| Challenge: | Existing RAG frameworks either indiscriminately perform retrieval or rely on rigid single-label classifiers to select retrieval methods. |
| Approach: | They propose a framework that dynamically selects the most suitable retrieval strategy based on query complexity. |
| Outcome: | The proposed framework achieves state-of-the-art results on multiple single-hop and multi-hop datasets while reducing retrieval costs. |
FlexQuant: A Flexible and Efficient Dynamic Precision Switching Framework for LLM Quantization (2025.findings-emnlp)
Copied to clipboard
Fangxin Liu, Zongwu Wang, Jinhong Xia, Junping Zhao, Shouren Zhao, Jinjin Li, Jian Liu, Li Jiang, Haibing Guan
| Challenge: | Existing methods for quantization of large language models struggle to adapt to dynamic workloads. |
| Approach: | a new framework optimizes the trade-off between inference speed and accuracy . FlexQuant enables fine-grained, layer-wise mixed-precision quantization . |
| Outcome: | a new framework optimizes the trade-off between inference speed and accuracy . it achieves a 1.3 speedup across diverse language tasks with negligible accuracy loss . |
Learning Global Controller in Latent Space for Parameter-Efficient Fine-Tuning (2024.acl-long)
Copied to clipboard
Zeqi Tan, Yongliang Shen, Xiaoxia Cheng, Chang Zong, Wenqi Zhang, Jian Shao, Weiming Lu, Yueting Zhuang
| Challenge: | Large language models (LLMs) have shown remarkable performance, but their training costs are exorbitant. |
| Approach: | They propose a parameter-efficient method for exploring optimal solutions within latent space by using latent units to extract input representations from LLMs. |
| Outcome: | The proposed method improves performance on a range of natural language processing tasks. |
Outside Computation with Superior Functions (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing generalized inside-outside algorithm would violate Strong Exponential Time Hypothesis (SETH) |
| Approach: | They propose a framework for efficient outside computation that would yield a sub-exponential time algorithm for SAT, violating the Strong Exponential Time Hypothesis (SETH). |
| Outcome: | The proposed algorithm would violate the Strong Exponential Time Hypothesis (SETH) . |
Improving Model Evaluation using SMART Filtering of Benchmark Datasets (2025.naacl-long)
Copied to clipboard
| Challenge: | Creating high quality human-annotated datasets is difficult due to dataset saturation. |
| Approach: | They propose a method to filter a subset of test examples from existing benchmarks by removing less informative and lower quality examples. |
| Outcome: | The proposed method reduces dataset size by 48% while increasing Pearson correlation with rankings from ChatBot Arena. |
Graph-based Dependency Parsing with Graph Neural Networks (P19-1)
Copied to clipboard
| Challenge: | In graph-based dependency parsers, learning representations is gaining in importance, and we use graph neural networks to learn the representations. |
| Approach: | They propose to use graph neural networks to learn dependency tree nodes and propose to add a new aggregation function to the system. |
| Outcome: | The proposed model achieves the best UAS and LAS on PTB (96.0%, 94.3%) without using external resources. |
Constrained Decoding with Speculative Lookaheads (2025.naacl-long)
Copied to clipboard
Nishanth Sridhar Nakshatri, Shamik Roy, Rajarshi Das, Suthee Chaidaroon, Leonid Boytsov, Rashmi Gangadharaiah
| Challenge: | Constrained decoding with lookahead heuristics is effective for aligning LLM generations to human preferences, but the extensive lookaheaded roll-out operations for each generated token make it prohibitively expensive. |
| Approach: | They propose a technique that uses lookaheads to align LLMs to human preferences . they propose 2.2x to 12.15x speedup over greedy decoding . |
| Outcome: | The proposed technique achieves 2.2x to 12.15x speedup over greedy decoding without significant performance reduction. |
QSpec: Speculative Decoding with Complementary Quantization Schemes (2025.emnlp-main)
Copied to clipboard
| Challenge: | Quantization is widely adopted to accelerate inference and reduce memory consumption in large language models. |
| Approach: | They propose a quantization paradigm that decouples efficiency from quality by integrating two complementary schemes via speculative decoding. |
| Outcome: | The proposed approach achieves 1.64x speedup without quality degradation and outperforms state-of-the-art speculative decoding methods by 1.55x in batched settings. |
COIL: Revisit Exact Lexical Match in Information Retrieval with Contextualized Inverted List (2021.naacl-main)
Copied to clipboard
| Challenge: | Recent neural IR models shift towards soft matching all query document terms, but they lose the computation efficiency of exact match systems. |
| Approach: | They propose a contextualized exact match retrieval architecture where scoring is based on overlapping query document tokens’ contextualized representations. |
| Outcome: | The proposed architecture outperforms classical lexical retrieval systems and state-of-the-art deep language models with smaller latency. |
Efficient Nearest Neighbor based Uncertainty Estimation for Natural Language Processing Tasks (2025.findings-naacl)
Copied to clipboard
| Challenge: | Existing methods for uncertainty estimation are inadequate for safety-critical applications. |
| Approach: | They propose a method that uses the distances from neighbors and the ratio of labels in neighbors to estimate uncertainty. |
| Outcome: | The proposed method outperforms baseline and density-based methods in calibration and uncertainty metrics. |
SLIM: Let LLM Learn More and Forget Less with Soft LoRA and Identity Mixture (2025.naacl-long)
Copied to clipboard
| Challenge: | balancing the training budget, downstream performance, and general capabilities of large language models remains a challenge in many applications. |
| Approach: | They propose a mixture of expert framework based on Soft LoRA and Identity Mixture . SLIM allows dynamic routing between LoRA adapters and identity layers . |
| Outcome: | The proposed framework reduces training cost while maintaining general capabilities . it can be open-sourced upon publication. |
Sparse Teachers Can Be Dense with Knowledge (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for transferring knowledge from a teacher of large scale to a student of smaller scale are limiting in overall knowledgeableness. |
| Approach: | They propose a sparse teacher trick to remove over-parameterized teachers that produce student-unfriendly knowledge and thus limit overall knowledgeableness. |
| Outcome: | The proposed trick removes the parameters that result in student-unfriendliness and leads to compelling performance in comparison with baselines. |
From Numbers to Narratives: Efficient Language Model-Based Detection for Safety-Critical Minority Classes (2026.findings-eacl)
Copied to clipboard
| Challenge: | Tabular datasets with high overall accuracy and poor performance on minority classes are often misaligned . a numbers to narratives framework improves overall accuracy by up to 22.43% in five of six datasets while maintaining computational feasibility. |
| Approach: | They propose a number-to-narrative framework that transforms tabular data into contextually rich descriptions. |
| Outcome: | The proposed framework achieves superior minority class F1-scores in five of six datasets. |
MegaAgent: A Large-Scale Autonomous LLM-based Multi-Agent System Without Predefined SOPs (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing multi-agent systems lack agent coordination and rely on predefined procedures . existing systems lack adaptive task coordination when task is big and complex . |
| Approach: | They propose a large-scale autonomous LLM-based multi-agent system that generates agents based on task complexity and enables dynamic task decomposition, parallel execution, efficient communication and comprehensive system monitoring. |
| Outcome: | The proposed system outperforms existing systems in task completion efficiency and scalability. |
MoDification: Mixture of Depths Made Easy (2025.naacl-long)
Copied to clipboard
Chen Zhang, Meizhi Zhong, Qimeng Wang, Xuantao Lu, Zheyu Ye, Chengqiang Lu, Yan Gao, Yao Hu, Kehai Chen, Min Zhang, Dawei Song
| Challenge: | Long-context efficiency is a trending topic in large language model (LLM) serving. |
| Approach: | They propose a method to combine long-context efficiency and mixture of depths to bring down both latency and memory. |
| Outcome: | The proposed method achieves 1.2 speedup in latency and 1.8 reduction in memory compared to original LLMs especially in long-context applications. |
KV Cache Compression, But What Must We Give in Return? A Comprehensive Benchmark of Long Context Capable Approaches (2024.findings-emnlp)
Copied to clipboard
Jiayi Yuan, Hongyi Liu, Shaochen Zhong, Yu-Neng Chuang, Songchen Li, Guanchu Wang, Duy Le, Hongye Jin, Vipin Chaudhary, Zhaozhuo Xu, Zirui Liu, Xia Hu
| Challenge: | Long context capability is a crucial competency for large language models as it mitigates the human struggle to digest long-form texts. |
| Approach: | They propose to evaluate 10+ state-of-the-art approaches for long context-capable LLMs. |
| Outcome: | The proposed methods are compared against 10+ state-of-the-art approaches across seven categories of long context tasks. |
Sparse Brains are Also Adaptive Brains: Cognitive-Load-Aware Dynamic Activation for LLMs (2026.findings-eacl)
Copied to clipboard
| Challenge: | Existing sparsity methods lack adaptivity to contextual or model structural demands or incur prohibitive computational overhead. |
| Approach: | They propose a Cognitive-Load-Aware Dynamic Activation framework that synergizes statistical sparsity with semantic adaptability. |
| Outcome: | The proposed framework achieves 20% average speedup with less than 2% accuracy degradation outperforming Griffin and TT. |
Cost-Optimal Grouped-Query Attention for Long-Context Modeling (2025.emnlp-main)
Copied to clipboard
Yingfa Chen, Yutong Wu, Chenyang Song, Zhen Leng Thai, Xingyu Shen, Xu Han, Zhiyuan Liu, Maosong Sun
| Challenge: | Current GQA configurations overlook how context length influences inference cost . |
| Approach: | They propose a recipe for deriving cost-optimal GQA configurations that decouple the total head size from the hidden size and allow more flexible control over attention FLOPs. |
| Outcome: | The proposed configurations reduce memory usage and FLOPs by more than 50% compared to Llama-3's GQA, with *no degradation in model capabilities*. |
Mirroring Users: Towards Building Preference-aligned User Simulator with User Feedback in Recommendation (2026.acl-long)
Copied to clipboard
| Challenge: | Large Language Models lack specific task alignment and large-scale simulations are challenging due to their ambiguity, noise and massive volume. |
| Approach: | They propose a framework that leverages user feedback in RSs with advanced LLM capabilities to generate high-quality simulation data. |
| Outcome: | The proposed framework boosts the alignment with human preferences and in-domain reasoning capabilities of the fine-tuned LLMs. |
Adaptable Adapters (2022.naacl-main)
Copied to clipboard
| Challenge: | Existing work uses the same adapter architecture for every dataset regardless of the properties of the dataset or the amount of training data. |
| Approach: | They propose to use adaptable adapters to finetune lightweight neural network layers on top of pretrained weights. |
| Outcome: | The proposed adapters achieve on-par performances with the standard adapter architecture while using a considerably smaller number of adapter layers. |
In Defense of Structural Sparse Adapters for Concurrent LLM Serving (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models (LLMs) require adapters to fine tune performance without extensive retraining. |
| Approach: | They propose a system that uses structurally sparse adapters to serve LLMs with multiple structurally-sparse axons. |
| Outcome: | The proposed system achieves 2.12 speedup over low-rank adapters on 96 adapters with a single GPU. |
EcoLANG: Efficient and Effective Agent Communication Language Induction for Social Simulation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models (LLMs) have demonstrated impressive ability to role-play humans and replicate complex social dynamics. |
| Approach: | They propose an efficient agent communication language induction for social simulations that reduces token consumption by over 20%. |
| Outcome: | The proposed model reduces token consumption by over 20% while preserving human language. |
MoLA: MoE LoRA with Layer-wise Expert Allocation (2025.findings-naacl)
Copied to clipboard
Chongyang Gao, Kezhen Chen, Jinmeng Rao, Ruibo Liu, Baochen Sun, Yawen Zhang, Daiyi Peng, Xiaoyuan Guo, Vs Subrahmanian
| Challenge: | Recent efforts to integrate low-rank adaptation (LoRA) with the Mixture-of-Experts (MoE) have achieved performance comparable to full-parameter fine-tuning by tuning much fewer parameters. |
| Approach: | They propose a parameter-efficient MoE method for low-rank adaptation with the Mixture-of-Experts (MoE) they use layers of LoRA experts to allocate more LoRA expert to middle layers . |
| Outcome: | The proposed method outperforms baseline models on six well-known NLP and commonsense QA benchmarks on LLAMA-2, Mistral, and Gemma. |
MaPPER: Multimodal Prior-guided Parameter Efficient Tuning for Referring Expression Comprehension (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for Referring Expression Comprehension (REC) lack specific domain abilities for precise local visual perception and visual-language alignment. |
| Approach: | They propose a framework for Parameter-Efficient Transfer Learning to localize a visual region via natural language using a prior-guided prior. |
| Outcome: | The proposed framework achieves the best accuracy compared to the current methods with only 1.41% tunable backbone parameters. |
POP: Prefill-Only Pruning for Efficient Large Model Inference (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing structured pruning methods suffer from significant accuracy degradation . Existing pruning methods are expensive and require specialized hardware and kernels to perform . |
| Approach: | They propose a stage-agnostic pruning approach that overlooks asymmetric roles between prefill and decode stages. |
| Outcome: | The proposed pruning approach achieves 1.37 speedup in prefill latency with minimal performance loss. |
TReX: Tokenizer Regression for Optimal Data Mixture (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing approaches to train and inference tokenizers rely on heuristics or large-scale searches to determine optimal data mixtures. |
| Approach: | They propose a regression-based framework that efficiently predicts the optimal data mixture for tokenizer training. |
| Outcome: | The proposed model outperforms mixtures based on LLaMA3 and uniform distributions by up to 12% in both in- and out-of-distribution compression efficiency. |
Efficient Second-Order TreeCRF for Neural Dependency Parsing (2020.acl-main)
Copied to clipboard
| Challenge: | In the deep learning (DL) era, dependency parsing models are extremely simplified with little hurt on performance thanks to the remarkable capability of multi-layer BiLSTMs in context representation. |
| Approach: | They propose to extend the biaffine parser to a second-order TreeCRF extension to reduce the complexity of the inside-outside algorithm. |
| Outcome: | The proposed extension can be used to batchify the inside and Viterbi algorithms and avoid the complex outside algorithm via efficient back-propagation. |
Beyond Distillation: Task-level Mixture-of-Experts for Efficient Inference (2021.findings-emnlp)
Copied to clipboard
Sneha Kudugunta, Yanping Huang, Ankur Bapna, Maxim Krikun, Dmitry Lepikhin, Minh-Thang Luong, Orhan Firat
| Challenge: | Sparse Mixture-of-Experts (MoE) is a successful approach for scaling multilingual translation models to billions of parameters without a proportional increase in training computation. |
| Approach: | They propose to use a task-level routing approach to extract smaller, ready-to-deploy sub-networks from large sparse models by ignoring distillation. |
| Outcome: | Experiments on WMT and a web-scale dataset show that task-level routing outperforms token-level MoE models by +1.0 BLEU on average across 30 language pairs. |
LATENTLOGIC: Learning Logic Rules in Latent Space over Knowledge Graphs (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for learning logic rules for knowledge graph reasoning face limitations such as searching in vast search space and inefficient optimization. |
| Approach: | They propose a framework to efficiently mine logic rules by controllable generation in the latent space by a pre-trained VAE and a discriminator. |
| Outcome: | The proposed framework efficiently mines logic rules by controllable generation in the latent space. |
MAS-Bench: A Unified Benchmark for Shortcut-Augmented Hybrid Mobile GUI Agents (2026.acl-long)
Copied to clipboard
Pengxiang Zhao, Guangyi Liu, Yaozhen Liang, Weiqing He, Zhengxi Lu, WenHao Wang, Yuehao Huang, Yuxiang Chai, Zhaolu Kang, Yaxuan Guo, Hao Wang, Kexin Zhang, Liang Liu, Yong Liu
| Challenge: | Shortcuts such as APIs and deep-links have emerged as efficient complements to flexible GUI operations, but systematic evaluation of GUI–shortcut hybrid agents remains underexplored. |
| Approach: | They propose a benchmark that evaluates GUI-shortcut hybrid agents with a specific focus on the mobile domain. |
| Outcome: | MAS-Bench evaluates agent's ability to generate shortcuts by discovering and creating reusable, low-cost workflows. |
Tiny Budgets, Big Gains: Parameter Placement Strategy in Parameter Super-Efficient Fine-Tuning (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods such as LoRA and VeRA use memory-efficient methods to fine-tune large language models. |
| Approach: | They propose a method that uses only 1–5% of the standard LoRA parameters and achieves state-of-the-art performance across a wide range of tasks. |
| Outcome: | The proposed method achieves state-of-the-art performance across a wide range of tasks using only 1–5% of the standard LoRA parameters. |
Efficient Learned Data Compression via Dual-Stream Feature Decoupling (2026.acl-long)
Copied to clipboard
| Challenge: | Learned data compression has achieved superior compression ratios, but balancing precise probability modeling with system efficiency remains challenging. |
| Approach: | They propose a Dual-Stream Multi-Scale Decoupler that disentangles local and global contexts to replace deep serial processing with shallow parallel streams. |
| Outcome: | The proposed method achieves state-of-the-art performance in both compression ratio and throughput while maintaining the lowest latency and memory usage. |
Efficient Algorithms for Recognizing Weighted Tree-Adjoining Languages (2023.emnlp-main)
Copied to clipboard
| Challenge: | a class of tree-adjoining languages can be characterized by various two-level formalisms controlled by semiring-weighted CFGs and PDAs. |
| Approach: | They propose semiring-weighted versions of controllable CFGs and PDAs . they also introduce a WPDA normal form that is analogous to Chomsky's normal form for CFG . |
| Outcome: | The proposed algorithms are more time-efficient than the previous ones for LIG, PAA, and EPDA. |
OSC: Cognitive Orchestration through Dynamic Knowledge Alignment in Multi-Agent LLM Collaboration (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Prior work has advanced agent selection and result aggregation, efficient linguistic interactions for deep collaboration among expert agents remain a critical bottleneck. |
| Approach: | They propose a knowledge-aware adaptive collaboration framework to enhance cognitive synergy in multi-agent systems with large language models. |
| Outcome: | The proposed framework improves synergy between agents and language models by enabling agents to dynamically perceive their collaborators’ cognitive states. |
Beyond Accuracy: Unveiling Inefficiency Patterns in Tool-Integrated Reasoning (2026.acl-long)
Copied to clipboard
| Challenge: | Tool-Integrated Reasoning (TIR) is a tool that can be used to solve complex tasks. |
| Approach: | They propose a hardware-aware TIR-efficiency metric that unifies internal reasoning and external tool-use costs while explicitly accounting for non-reusable KV-Cache and long-tool-response scenarios. |
| Outcome: | The proposed metric explains wall-clock latency significantly better than token-count metric in a simulated high-concurrency industrial setting. |
A Unified Encoding of Structures in Transition Systems (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to encode dynamic structures only encode partial information of structures. |
| Approach: | They propose a new attention-based encoder unifying all structures in a transition system. |
| Outcome: | The proposed method significantly improves the test speed and achieves the best transition-based model. |
RMoA: Optimizing Mixture-of-Agents through Diversity Maximization and Residual Compensation (2025.findings-acl)
Copied to clipboard
| Challenge: | Multi-agent systems based on large language models are limited by high computational overhead, information loss, and robustness. |
| Approach: | They propose a Residual Mixture-of-Agents (RMoA) that integrates residual connections to optimize efficiency and reliability. |
| Outcome: | The proposed model achieves state-of-the-art performance on benchmarks of alignment, mathematical reasoning, code generation, and multitasking understanding, while significantly reducing computational overhead. |
Dynamic Scaling of Unit Tests for Code Reward Modeling (2025.acl-long)
Copied to clipboard
| Challenge: | Existing large language models struggle to produce accurate responses on the first attempt for complex reasoning tasks like code generation. |
| Approach: | They propose a lightweight yet effective unit test generator that scales unit tests based on problem difficulty. |
| Outcome: | The proposed approach significantly improves performance on three benchmarks. |
OS-Kairos: Adaptive Interaction for MLLM-Powered GUI Agents (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing efforts to build GUI agents focused on the autonomous mode have failed to address the problem of over-execution. |
| Approach: | They propose an adaptive GUI agent that predicts confidence levels at each interaction step and elicits adaptive interaction. |
| Outcome: | The proposed GUI agent outperforms existing models on a complex dataset and on established benchmarks. |
Train Flat, Then Compress: Sharpness-Aware Minimization Learns More Compressible Models (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Recent advances in hardware, modeling, and optimization for deep neural networks have led to improvements in memory and inference efficiency. |
| Approach: | They propose to combine sharpness-aware minimization with various model compression methods to improve model compressibility. |
| Outcome: | Empirically, optimizing for flatter minima leads to greater compressibility of parameters compared to vanilla Adam when fine-tuning BERT models, with little to no loss in accuracy on the GLUE text classification and SQuAD question answering benchmarks. |
LoRA-drop: Efficient LoRA Parameter Pruning based on Output Evaluation (2025.coling-main)
Copied to clipboard
| Challenge: | Low-Rank Adaptation (LoRA) is currently the most commonly used PEFT method for fine-tuning models with billions of parameters. |
| Approach: | They propose to use low-rank Adaptation to evaluate LoRA parameter features and then retain LoRA for important layers and the other layers share the same LoRA. |
| Outcome: | The proposed method achieves comparable performance to full fine-tuning and LoRA while retaining 50% of the LoRA parameters on average. |
LycheeCluster: Efficient Long-Context Inference with Structure-Aware Chunking and Hierarchical KV Indexing (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing retrieval-based methods compromise semantic integrity through fixed-size chunking and suffer from inefficient linear scanning. |
| Approach: | They propose a method that preserves local semantic coherence through boundary-aware chunking and constructs a recursive hierarchical index rooted in the triangle inequality. |
| Outcome: | The proposed method achieves 3.6 end-to-end inference speedup with negligible degradation in model performance. |
Efficient LLM Comparative Assessment: A Product of Experts Framework for Pairwise Comparisons (2024.emnlp-main)
Copied to clipboard
| Challenge: | LLM-as-a-judge approaches are effective but cost scales quadratically with number of candidates, which has practical limitations. |
| Approach: | They propose a Product of Expert (PoE) framework for efficient LLM Comparative Assessment where individual comparisons are considered experts that provide information on a pair’s score difference. |
| Outcome: | The proposed framework can generate score predictions that correlate well with human judgements on multiple NLG tasks with as few as 2% of comparisons. |
Utility is in the Eye of the User: A Critique of NLP Leaderboards (2020.emnlp-main)
Copied to clipboard
| Challenge: | Performance-based evaluation has been at the expense of other attributes valued by the NLP community, such as compactness and energy efficiency. |
| Approach: | They propose to frame both the leaderboard and NLP practitioners as consumers and the benefit they get from a model as its utility to them. |
| Outcome: | The proposed model size and energy efficiency benchmarks have been successful in driving the creation of more accurate models, but have been at the expense of other attributes valued by the NLP community. |
EcoTune: Token-Efficient Multi-Fidelity Hyperparameter Optimization for Large Language Model Inference (2025.emnlp-main)
Copied to clipboard
| Challenge: | Multi-fidelity optimization improves inference performance with low-fidelity evaluations, but its static scheduling strategies ignore token consumption, leading to high costs. |
| Approach: | They propose a token-efficient multi-fidelity optimization method that optimizes for inference hyperparameters . their method reduces token consumption by over 80% while maintaining or surpassing performance . |
| Outcome: | The proposed method improves over the HELM leaderboard by 7.1%, 24.3%, 21.9%, and 4.6% while maintaining or surpassing performance. |
Decoding in Latent Spaces for Efficient Inference in LLM-based Recommendation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Light Latent-space Decoding (L2D) is an efficient and efficient latent- space decoding method. |
| Approach: | They propose to bypass language-space decoding by matching candidate items with LLM's internal thought representations in the latent space. |
| Outcome: | The proposed method is 10x faster than language-space decoding while maintaining or enhancing performance. |
Scaling LLM Inference Efficiently with Optimized Sample Compute Allocation (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing methods to optimize sample allocations for large language models fail to account for the optimal sampling configuration. |
| Approach: | They propose an algorithm that optimizes sample allocation by finding an optimal mix of different inference configurations. |
| Outcome: | The proposed algorithm achieves better accuracy on SWE-Bench with 3x less compute than the default configuration. |
LongHeads: Multi-Head Attention is Secretly a Long Context Processor (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models struggle to process lengthy inputs due to limited length generalization and attention’s quadratic computational demands. |
| Approach: | They propose a training-free framework that allows each head to attend to important context chunks instead of allowing each head a full sentence . |
| Outcome: | The proposed framework unlocks multi-head attention's untapped potential by allowing each head to attend to important context chunks instead of the full sentence. |
Enhancing Cross-Tokenizer Knowledge Distillation with Contextual Dynamical Mapping (2025.findings-acl)
Copied to clipboard
| Challenge: | Knowledge distillation (KD) approaches focus on homogeneous architectures with identical tokenizers, constraining their applicability in cross-architecture scenarios. |
| Approach: | They propose a framework that uses contextual information to enhance sequence alignment precision and dynamically improves vocabulary mapping. |
| Outcome: | The proposed framework shows significant advantages over existing methods for model compression . it can be used across multiple model families and across multiple benchmarks . |
CORG: Generating Answers from Complex, Interrelated Contexts (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing approaches to analyzing knowledge in a corpus often focus on single factors in isolation. |
| Approach: | They propose a framework that organizes multiple contexts into independently processed groups . they classify these relationships into distracting, ambiguous, counterfactual, and duplicated . |
| Outcome: | The proposed framework outperforms existing grouping methods and single-context approaches. |
AnchorAttention: Difference-Aware Sparse Attention with Stripe Granularity (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for large language models with extended context lengths face significant computational challenges during the prefill phase. |
| Approach: | They propose a difference-aware, dynamic sparse attention mechanism that efficiently identifies critical attention regions at a finer stripe granularity while adapting to global contextual information. |
| Outcome: | The proposed model achieves a speedup of 1.44 while maintaining higher recall rates. |
Tiny-Attention Adapter: Contexts Are More Important Than the Number of Parameters (2022.emnlp-main)
Copied to clipboard
| Challenge: | Adapter-tuning is a paradigm that transfers a pretrained language model to downstream tasks . Previously proposed adapters are all feed-forward neural networks . |
| Approach: | They propose to use tiny-attention attention with extremely small per-head dimensionality as adapters to modify hidden states at each position . they propose to average multiple attention heads' weights during deployment to reduce its inference computation cost. |
| Outcome: | The proposed adapter outperforms other adapter-tuning methods on the GLUE benchmark . it uses attention with extremely small per-head dimensionality to modify hidden states . |
IIET: Efficient Numerical Transformer via Implicit Iterative Euler Method (2025.emnlp-main)
Copied to clipboard
Xinyu Liu, Bei Li, Jiahao Liu, Junhao Ruan, Kechen Jiao, Hongyin Tang, Jingang Wang, Tong Xiao, JingBo Zhu
| Challenge: | High-order numerical methods enhance performance in tasks like NLP but introduce a performance-efficiency trade-off due to increased computational overhead. |
| Approach: | They propose an iterative implicit Euler Transformer which simplifies high-order numerical methods by iterating implicit Eule. |
| Outcome: | The proposed method improves accuracy and reduces inference overhead by 55% while retaining 99.4% of the original task accuracy. |
Mitigating Negative Interference in Multilingual Knowledge Editing through Null-Space Constraints (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing monolingual knowledge editing methods are expensive and require multiple models to maintain factual consistency. |
| Approach: | They propose a null-space constrained framework to precisely isolate language-specific knowledge updates that can be mapped onto other languages’ subspaces. |
| Outcome: | The proposed framework can project parameter updates for each language onto the orthogonal complement of other languages’ subspaces while preserving multilingual generalization capabilities. |
TT-SI: Self-Improving LLM Agents with Test-Time Training (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for language model fine-tuning are expensive and inefficient . existing methods rarely assess whether a training sample provides novel information . |
| Approach: | They propose a test-time self-improvement algorithm that generates a sample that model struggles with . they also explore Test-Time Distillation, which leverages 'stronger supervisors' |
| Outcome: | The proposed algorithm improves performance with +5.48% absolute accuracy gain on average across benchmarks. |
Can LLMs Help You at Work? A Sandbox for Evaluating LLM Agents in Enterprise Environments (2025.emnlp-main)
Copied to clipboard
| Challenge: | Enterprise systems are crucial for enhancing productivity and strategic growth, but data is fragmented across multiple sources and access controls are complex. |
| Approach: | They propose a benchmark that simulates enterprise settings with 500 diverse tasks . they show that even the most capable models achieve only 41.8% task completion . |
| Outcome: | The proposed benchmark shows that even the most capable models achieve only 41.8% task completion. |
Tied-LoRA: Enhancing parameter efficiency of LoRA with Weight Tying (2024.naacl-long)
Copied to clipboard
| Challenge: | a new paradigm for low-rank Adaptation (LoRA) uses weight tying and selective training to improve parameter efficiency. |
| Approach: | They propose a paradigm that uses weight tying and selective training to enhance parameter efficiency of Low-rank Adaptation. |
| Outcome: | The proposed paradigm achieves comparable performance to LoRA with reduced model complexity . the proposed paradigm can be used for a variety of tasks and languages . |
Towards Pareto-Efficient RLHF: Paying Attention to a Few High-Reward Samples with Reward Dropout (2024.findings-emnlp)
Copied to clipboard
| Challenge: | RLHF is a bi-objective problem that has the nature of a Pareto optimization . reward dropout is generalizable and most effective with non-pretrained target models . |
| Approach: | They propose a method that guarantees a Pareto improvement by leveraging reinforcement learning to fine-tune language models. |
| Outcome: | The proposed method guarantees a Pareto improvement on two benchmark datasets . it is generalizable and most effective with non-pretrained target models, saving the effort of pretraining. |
Stronger, Lighter, Better: Towards Life-Long Attribute Value Extraction for E-Commerce Products (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing models for attribute value extraction struggle for parameter efficiency and reliability due to data contamination and catastrophic forgetting. |
| Approach: | They propose to decouple product type and attribute to promote de-contamination and parameter efficiency while scaling up. |
| Outcome: | The proposed model achieves state-of-the-art performance with affordable parameter size, least historical knowledge forgetting, and greatest robustness against noises. |
Sherry: Hardware-Efficient 1.25-Bit Ternary Quantization via Fine-grained Sparsification (2026.acl-long)
Copied to clipboard
| Challenge: | ternary quantization is a powerful solution for resource-constrained edge devices . current implementations suffer from a fundamental misalignment with commodity hardware . |
| Approach: | They propose a hardware-efficient ternary quantization framework that packs weights into five bits to restore power-of-two alignment. |
| Outcome: | The proposed framework reduces weights to -1, 0, +1 while preserving power-of-two alignment. |
LEDGER: Scaling Agentic Document Editing with Dependency-aware Graph Retrieval (2026.findings-acl)
Copied to clipboard
| Challenge: | Document editing requires full-context awareness of dependencies, but processing entire documents for each edit incurs prohibitive token costs and latency. |
| Approach: | a framework that constructs lightweight dependency graphs captures semantic relationships and structural hierarchies across document elements is proposed for agentic document editing . a scaLing agentic agentic framework is based on a dependency graph framework that captures dependencies and refactors function dependencies. |
| Outcome: | a new framework achieves 76 consistency versus 56 baseline while reducing token usage by 85 . the framework is based on a framework that captures semantic relationships and structural hierarchies across document elements . it can be used to improve document consistency, but it also reduces token costs and latency . |
DB-LLM: Accurate Dual-Binarization for Efficient LLMs (2024.findings-acl)
Copied to clipboard
Hong Chen, Chengtao Lv, Liang Ding, Haotong Qin, Xiabin Zhou, Yifu Ding, Xuebo Liu, Min Zhang, Jinyang Guo, Xianglong Liu, Dacheng Tao
| Challenge: | Existing methods for ultra-low bit quantization cause severe accuracy drops . a novel Dual-Binarization method is proposed for efficient Large Language Models . |
| Approach: | They propose a Dual-Binarization method that takes 2-bit-width and binarization into account . they propose DB-LLM, which uses a 2-bit binarized weighted model to represent weights efficiently . |
| Outcome: | The proposed method surpasses the current State-of-the-Art in ultra-low bit quantization and achieves 20% reduction in computational consumption compared to the SOTA method under the same bit-width. |
Beyond Pedagogical Principles: Multi-Horizon Preference Optimization for Efficient Socratic Tutoring (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for developing LLMs are constrained by static data or sparse reward signals in online settings. |
| Approach: | They propose a framework that iteratively refines tutor agents using a multi-horizon reward function within a dynamic teacher-student simulation environment. |
| Outcome: | The proposed framework improves model performance and balances principles and effectiveness compared to baselines. |
On Efficient Retrieval of Top Similarity Vectors (D19-1)
Copied to clipboard
| Challenge: | Existing representation learning methods such as Word2vec represent word embeddings in the semantic space. |
| Approach: | They propose an efficient method for searching vectors via a non-metric matching function: inner product. |
| Outcome: | Experiments on data representations learned for different machine learning tasks show the proposed method outperforms existing methods. |
MoQAE: Mixed-Precision Quantization for Long-Context LLM Inference via Mixture of Quantization-Aware Experts (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to optimize large language models for long-context inference are inefficient and consume memory. |
| Approach: | They propose a mixed-precision quantization method via mixture of experts that inputs tokens into router chunk by chunk to reduce inference overhead. |
| Outcome: | The proposed method outperforms state-of-the-art KV cache quantization methods on multiple benchmark datasets. |
Fast Forwarding Low-Rank Training (2024.emnlp-main)
Copied to clipboard
| Challenge: | Modern optimizers provide a spectacular array of tweaks to stabilize training trajectories and accelerate Stochastic Gradient Descent (SGD). |
| Approach: | They propose a fast-forward approach to accelerate large segments of SGD training . they alternate between Adam SGD for burn-in and accelerating by line search . |
| Outcome: | The proposed approach speeds up training without compromising model performance. |
Pre3: Enabling Deterministic Pushdown Automata for Faster Structured LLM Generation (2025.acl-long)
Copied to clipboard
Junyi Chen, Shihao Bai, Zaijun Wang, Siyu Wu, Chuheng Du, Hailong Yang, Ruihao Gong, Shengzhong Liu, Fan Wu, Guihai Chen
| Challenge: | Existing methods for structured generation of outputs are inefficient under large inference batches. |
| Approach: | They propose a new LLM-based method that parses LR(1) grammars into a pushdown automaton and exploits deterministic pushdown automation to optimize the constrained LLM decoding efficiency. |
| Outcome: | The proposed method improves time per output token (TPOT) by 40% and throughput by 36% . |
RouteMoA: Dynamic Routing without Pre-Inference Boosts Efficient Mixture-of-Agents (2026.acl-long)
Copied to clipboard
Jize Wang, Han Wu, Zhiyuan You, Yiming Song, Yijun Wang, Zifei Shan, Yining Li, Songyang Zhang, Xinyi Le, Cailian Chen, Xinping Guan, Dacheng Tao
| Challenge: | Existing methods for mixing-of-agents (MoA) lack model selection criteria and struggle with large model pools. |
| Approach: | They propose a mixture-of-agents framework with dynamic routing that uses a lightweight scorer to perform initial screening and refines the model scores through self- and cross-assessment. |
| Outcome: | The proposed framework outperforms existing methods for large model pools and tasks . it reduces cost by 89.8% and latency by 63.6% in the large-scale model pool. |
Text or Pixels? Evaluating Efficiency and Understanding of LLMs with Visual Text Inputs (2025.findings-emnlp)
Copied to clipboard
| Challenge: | *visual text representations* are a practical and surprisingly effective form of input compression for decoder LLMs. |
| Approach: | They exploit visual representations to render long text inputs as a single image and provide it directly to the model. |
| Outcome: | The proposed method reduces token usage while preserving performance. |
Cache-Efficient Posterior Sampling for Reinforcement Learning with LLM-Derived Priors Across Discrete and Continuous Domains (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) are used as action proposers in reinforcement learning (RL) but they are expensive and require prohibitive computational costs. |
| Approach: | They propose a cache-efficient framework for Bayesian RL that leverages large language models as action proposers and optimizes meta-learning based on policy performance to enable efficient inference across text-based games and robotic control tasks. |
| Outcome: | The proposed framework achieves 3.8–4.7 reduction in LLM queries and 4.0–12.0 lower median latencies (85–93ms on consumer hardware) while retaining 96–98% of the uncached policy’s performance. |
WeightLoRA: Keep Only Necessary Adapters (2026.acl-long)
Copied to clipboard
| Challenge: | Low-rank adaptation (LoRA) adds trainable adapters to selected layers, but requires significant memory to train large models and intuition on which layers to add adapters. |
| Approach: | They propose a method which adds trainable adapters to selected layers . they compare weightLoRA with different adaptive approaches to reduce trainable parameters while maintaining consistent or even superior metric values. |
| Outcome: | The proposed method reduces the number of trainable parameters while maintaining the capability to obtain consistent or even superior metric values. |
Flashback: Memory Mechanism for Enhancing Memory Efficiency and Speed in Deep Sequential Models (2025.coling-main)
Copied to clipboard
| Challenge: | Existing deep sequential processing models have problems with memory degradation and inaccurate gradient backpropagation. |
| Approach: | They propose a Flashback property that preserves memory as an identity mapping until it is overwritten by a hidden state at a different time step. |
| Outcome: | The proposed model can be implemented in Transformers and Mamba, and it performs well. |
Improving the Efficiency of Grammatical Error Correction with Erroneous Span Detection and Correction (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to improve the efficiency of GEC are not efficient enough for GEC. |
| Approach: | They propose a language-independent approach to improve the efficiency of GEC by dividing the task into two subtasks: ESD and ESC. |
| Outcome: | The proposed approach performs comparably to conventional seq2seq approaches in English and Chinese GEC benchmarks with less than 50% time cost for inference. |
Predicting Prerequisite Relations for Unseen Concepts (2022.emnlp-main)
Copied to clipboard
| Challenge: | Concept prerequisite learning (CPL) is a task of building a concept graph by structuring open knowledge in prerequisite relations. |
| Approach: | They propose to use both content-based and graph-based models to build a concept graph by structuring open knowledge in prerequisite relations. |
| Outcome: | The proposed approach improves F1 scores by 10% on three public benchmarks. |
From Characters to Tokens: Dynamic Grouping with Hierarchical BPE (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Subword tokenization methods suffer from inefficiencies in representing rare words and require large embedding matrices. |
| Approach: | They propose a dynamic character grouping method that leverages existing BPE tokenization without adding additional models. |
| Outcome: | The proposed method matches or exceeds existing entropy- and whitespace-based patching strategies while maintaining a compact vocabulary. |
A Goal Without a Plan Is Just a Wish: Efficient and Effective Global Planner Training for Long-Horizon Agent Tasks (2026.acl-long)
Copied to clipboard
Shuzheng Si, Haozhe Zhao, Kangyang Luo, Gang Chen, Fanchao Qi, Minjia Zhang, Baobao Chang, Maosong Sun
| Challenge: | Recent advances in large language models (LLMs) have leapt from static chatbots to versatile agents that tackle complex tasks such as science experiments. |
| Approach: | They propose a plan-and-execute framework and propose 'EAGLET' to enhance the executor agent's planning abilities without human effort. |
| Outcome: | The proposed method outperforms existing methods on three long-horizon tasks and reduces training costs by 8 compared to baselines. |
Optima: Optimizing Effectiveness and Efficiency for LLM-Based Multi-Agent System (2025.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have emerged as powerful tools for a wide range of tasks, from * Equal Contribution. |
| Approach: | They propose a framework that enhances communication efficiency and task effectiveness in LLM-based multi-agent systems through training. |
| Outcome: | The proposed framework improves communication efficiency and task effectiveness on multi-agent tasks with 2.8x performance gain with less than 10% tokens on tasks requiring heavy information exchange. |
A Secure and Efficient Federated Learning Framework for NLP (2021.emnlp-main)
Copied to clipboard
Chenghong Wang, Jieren Deng, Xianrui Meng, Yijue Wang, Ji Li, Sheng Lin, Shuo Han, Fei Miao, Sanguthevar Rajasekaran, Caiwen Ding
| Challenge: | Existing FL frameworks require a trusted aggregator or require heavy-weight cryptographic primitives, which makes the performance significantly degraded. |
| Approach: | They propose a framework that is federated and efficient for NLP . they propose to eliminate the need for trusted entities and achieve better model accuracy . |
| Outcome: | The proposed framework achieves better model accuracy and model accuracy than existing FL frameworks. |
Rehearsal-free Continual Language Learning via Efficient Parameter Isolation (2023.acl-long)
Copied to clipboard
Zhicheng Wang, Yufang Liu, Tao Ji, Xiaoling Wang, Yuanbin Wu, Congcong Jiang, Ye Chao, Zhencong Han, Ling Wang, Xu Shao, Wenqiu Zeng
| Challenge: | Existing methods for learning continual tasks do not cache history data, which makes the problem more challenging. |
| Approach: | They propose a method that allocates a small portion of private parameters and learns them with a shared pre-trained model. |
| Outcome: | The proposed method is comparable to existing methods and comparable to those using historical data. |
Learning to Use Tools via Cooperative and Interactive Agents (2024.findings-emnlp)
Copied to clipboard
Zhengliang Shi, Shen Gao, Xiuyi Chen, Yue Feng, Lingyong Yan, Haibo Shi, Dawei Yin, Pengjie Ren, Suzan Verberne, Zhaochun Ren
| Challenge: | Existing methods for large language models (LLMs) use one agent to iterate and execute tools, but they suffer from performance degradation when addressing practical tasks. |
| Approach: | They propose a tool learning framework that coordinates three specialized agents for tool selection, tool execution, and action calibration separately. |
| Outcome: | The proposed framework outperforms baseline models on three datasets with 14% higher success rate. |
AgentInit: Initializing LLM-based Multi-Agent Systems via Diversity and Expertise Orchestration for Effective and Efficient Collaboration (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing MAS initialization methods do not fully account for the collaborative needs of the generated agents in subsequent stages. |
| Approach: | They propose to use a Natural Language to Format mechanism to optimize the structure of agent teams and incorporate a natural language to format mechanism to ensure consistency and standardization. |
| Outcome: | The proposed method outperforms state-of-the-art initialization methods and pre-defined strategies across various frameworks and tasks while reducing token consumption. |
Vector-Vector-Matrix Architecture: A Novel Hardware-Aware Framework for Low-Latency Inference in NLP Applications (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to improve accuracy of neural networks are slow due to computational complexity. |
| Approach: | They propose a vector-vector-matrix architecture which greatly reduces latency at inference time for NLP applications by a factor of four. |
| Outcome: | The proposed framework reduces the latency of sequence-to-sequence and Transformer models used for NMT by a factor of four. |
Communication-Efficient Desire Alignment for Proactive Embodied Human–Agent Interaction (2026.acl-long)
Copied to clipboard
| Challenge: | Effective real-world human–agent interactions are long-term and repeated. |
| Approach: | They propose a simulation that uses a proxy user with value-driven preferences and natural language behavior to evaluate how agents adapt to users across interactions and satisfy their desires. |
| Outcome: | HA-Desire, a home assistance simulation, shows that agents can adapt to user needs and provide proactive assistance within limited communication. |
Maximum Score Routing For Mixture-of-Experts (2025.findings-acl)
Copied to clipboard
| Challenge: | Traditional mixture-of-experts (MoE) networks impose an expert capacity constraint to ensure GPU-friendly computation. |
| Approach: | They propose a routing paradigm that dynamically allocates input tokens to top-k experts through differentiable sparse transformations, enabling scalable model capacity while preserving computational efficiency. |
| Outcome: | The proposed model achieves lower training losses and higher evaluation scores at equivalent FLOPs compared to constrained and unconstrained baselines. |
Agentic Episodic Control (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for reinforcement learning (RL) are limited by poor data efficiency and weak generalization. |
| Approach: | They propose a novel architecture that integrates large language models into episodic RL. |
| Outcome: | The proposed architecture achieves 2–6 higher data efficiency than baselines and is the only method to solve complex tasks like UnlockLocal with over 90% success. |
ModeX: Evaluator-Free Best-of-N Selection for Open-Ended Generation (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to selecting a single high-quality output from multiple generations are limited in their applicability and efficiency. |
| Approach: | They propose a method that generalizes majority voting to open-ended text generation . modeX leverages structural information across multiple generation paths to select a "modal" output . |
| Outcome: | The proposed framework outperforms standard single- and multi-path baselines in open-ended tasks. |
Disentangle to Decay: Linear Attention with Trainable Decay Factor (2025.coling-main)
Copied to clipboard
| Challenge: | Existing linear attention models use a decay factor based positional encoding (PE), but the decay factor is manually designed and non-trainable, limiting further optimization. |
| Approach: | They propose a PE-based positional encoding that disentangles decay factor into two parts to achieve further optimization and stable training. |
| Outcome: | The proposed model achieves stable training of decay factor and improves inference efficiency in normal context and extrapolation scenarios. |
CITADEL: Conditional Token Interaction via Dynamic Lexical Routing for Efficient and Effective Multi-Vector Retrieval (2023.acl-long)
Copied to clipboard
Minghan Li, Sheng-Chieh Lin, Barlas Oguz, Asish Ghoshal, Jimmy Lin, Yashar Mehdad, Wen-tau Yih, Xilun Chen
| Challenge: | Existing multi-vector retrieval methods are slower and require more space to store indices compared to their single-vektor counterparts. |
| Approach: | They propose a multi-vector retrieval method that uses dynamic lexical routing to route different token vectors to the predicted lexicals. |
| Outcome: | The proposed method achieves state-of-the-art performance on several benchmark datasets while being nearly 40 times faster than the current state-out-of the-art method. |
TL-CL: Task And Language Incremental Continual Learning (2024.emnlp-main)
Copied to clipboard
| Challenge: | a multilingual model is periodically updated to accommodate new tasks in previously learned languages or new languages for established tasks. |
| Approach: | They propose an adapter-based parameter-efficient fine-tuning strategy for continual learning in multilingual models. |
| Outcome: | The proposed approach outperforms other parameter-efficient approaches without access to historical data for replay. |
Graph-Structured Speculative Decoding (2024.findings-acl)
Copied to clipboard
Zhuocheng Gong, Jiahao Liu, Ziyue Wang, Pengfei Wu, Jingang Wang, Xunliang Cai, Dongyan Zhao, Rui Yan
| Challenge: | Speculative decoding is a promising technique to accelerate the inference of Large Language Models. |
| Approach: | They propose a method that uses a token graph to record multiple sequence hypotheses within a single draft stage. |
| Outcome: | The proposed method significantly accelerates the inference of Large Language Models (LLMs) it allows the LLM to choose from and select the longest sequence that meets its standards. |
RA2FD: Distilling Faithfulness into Efficient Dialogue Systems (2024.emnlp-main)
Copied to clipboard
| Challenge: | Retrieval Augmented Generation (RAG) is effective but inference inefficient, while Retrieral Free Generations (RFG) are more efficient but sacrifice faithfulness. |
| Approach: | They propose a retrieval-free model training scheme that uses a teacher-student framework to distill the faithfulness capacity of a student's knowledge-infused responses. |
| Outcome: | The proposed model surpasses the previous SOTA RFG model on knowledge-grounded dialogue datasets by an average of 33% while improving inference efficiency. |
AutoJudger: An Agent-Driven Framework for Efficient Benchmarking of MLLMs (2026.acl-long)
Copied to clipboard
| Challenge: | Evaluating multimodal large language models (MLLMs) is becoming increasingly expensive as benchmarks grow in scale and cross-modality complexity. |
| Approach: | They propose an adaptive evaluation framework for efficient benchmarking that treats evaluation as an interview-like process by keeping a hypothesized ability structure of the evaluated model and actively selecting the informative questions. |
| Outcome: | Experiments on four representative multimodal benchmarks show that **A2-Judger significantly improves sample efficiency while maintaining reliable evaluation results. |
Best-k Search Algorithm for Neural Text Generation (2023.acl-long)
Copied to clipboard
| Challenge: | Modern natural language generation paradigms require a decoding strategy to obtain quality sequences out of the model. |
| Approach: | They propose a deterministic search algorithm balancing quality and diversity . they investigate the vanilla best-first search algorithm and propose k-k search algorithm. |
| Outcome: | The proposed algorithm is parameter-free, lightweight, efficient, and easy-to-use. |
LLM-TOPLA: Efficient LLM Ensemble by Maximising Diversity (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Extensive evaluation of modern large language models shows performance gain over component LLMs. |
| Approach: | They propose a diversityoptimized LLM ensemble method with three unique properties . they introduce the focal diversity metric to capture diversityperformance correlation . |
| Outcome: | The proposed method outperforms the best-performing ensemble on four benchmarks. |
MedVerse: Efficient and Reliable Medical Reasoning via DAG-Structured Parallel Execution (2026.acl-long)
Copied to clipboard
Jianwen Chen, Xinyu Yang, Peng Xia, Arian Azarang, Yueh Z Lee, Gang Li, Hongtu Zhu, Yun Li, Beidi Chen, Huaxiu Yao
| Challenge: | Recent advances in large reasoning models have broadened the capabilities of medical artificial intelligence. |
| Approach: | They propose a reasoning framework for complex medical inference that reformulates medical reasoning as a parallelizable directed acyclic graph process based on Petri Net theory. |
| Outcome: | The proposed reasoning framework improves strong general-purpose LLMs by up to 8.9%. |
Craw4LLM: Efficient Web Crawling for LLM Pretraining (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing work discards over 90% of the raw data collected from web crawls, highlighting the inefficiency of current web crawlers in collecting LLM pretraining data. |
| Approach: | They propose a web crawling method that leverages the preference of LLMs as the priority score of the web crawler’s scheduler to obtain high-quality pretraining data. |
| Outcome: | The proposed method achieves high-quality pretraining data on a web graph containing 900 million webpages from a commercial search engine's index with just 21% URLs crawled. |
Active Domain Knowledge Acquisition with 100-Dollar Budget: Enhancing LLMs via Cost-Efficient, Expert-Involved Interaction in Sensitive Domains (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have demonstrated an impressive level of general knowledge, but often struggle in highly specialized domains due to the lack of expert knowledge. |
| Approach: | They propose a framework to actively engage domain experts within a fixed budget to enhance domain-specific LLMs. |
| Outcome: | The proposed framework improves LLMs in highly specialized domains while adhering to budget constraints. |
AdapterShare: Task Correlation Modeling with Adapter Differentiation (2022.emnlp-main)
Copied to clipboard
| Challenge: | AdapterShare is an adapter differentiation method to explicitly model the task correlation among multiple tasks. |
| Approach: | They propose an adapter differentiation method to explicitly model the task correlation among multiple tasks. |
| Outcome: | The proposed method achieves 1.90 points improvement on five dialogue understanding tasks and 2.33 points gain on NLU tasks. |
MULFE: A Multi-Level Benchmark for Free Text Model Editing (2024.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have impressive capabilities in comprehending human language and vast parametric knowledge obtained from large corpora. |
| Approach: | They propose a multi-level benchmark for free text model editing to bridge the gap . they categorize probe queries into three levels of generalization . |
| Outcome: | The proposed method improves the generalization performance of large langugae models. |
No Train but Gain: Language Arithmetic for training-free Language Adapters enhancement (2025.coling-main)
Copied to clipboard
| Challenge: | Modular deep learning is the most effective way to lift the curse of multilinguality. |
| Approach: | They propose a method which enables training-free post-processing to address this limitation by adding learning to the language adapters and transitioning the framework from a multi-task to a multiple language setup. |
| Outcome: | The proposed method consistently improves baselines with significant gains, especially in the most challenging case of zero-shot application. |
Moving Beyond Downstream Task Accuracy for Information Retrieval Benchmarking (2023.findings-acl)
Copied to clipboard
Keshav Santhanam, Jon Saad-Falcon, Martin Franz, Omar Khattab, Avi Sil, Radu Florian, Md Arafat Sultan, Salim Roukos, Matei Zaharia, Christopher Potts
| Challenge: | Neural information retrieval (IR) systems have progressed rapidly in recent years . many IR benchmarks focus on downstream task accuracy, concealing costs incurred . |
| Approach: | They propose to include efficiency considerations on IR benchmarks to help drive progress . eral et al. propose to incorporate query latency and cost budgets into evaluation . |
| Outcome: | a new study shows that the best IR system varies according to how efficiency considerations are chosen and weighed . the proposed benchmarks would allow for more thorough exploration of possible system designs . |
Turn Waste into Worth: Rectifying Top-k Router of MoE (2024.emnlp-main)
Copied to clipboard
Zhiyuan Zeng, Qipeng Guo, Zhaoye Fei, Zhangyue Yin, Yunhua Zhou, Linyang Li, Tianxiang Sun, Hang Yan, Dahua Lin, Xipeng Qiu
| Challenge: | Top-k router suffers from redundancy computation and memory costs due to unbalanced routing . some experts are overflow, where exceeding tokens are dropped, while others are empty, which are padded with zeros, negatively impacting model performance. |
| Approach: | They propose a top-k router that is unbalanced and uses a multi-gPU system to handle dropped tokens and padding. |
| Outcome: | The proposed model surpasses the top-1 router by 4.7% in terms of performance . the top-k router suffers from redundancy computation and memory costs . |
Precedent-Enhanced Legal Judgment Prediction with LLM and Domain-Model Collaboration (2023.emnlp-main)
Copied to clipboard
Yiquan Wu, Siying Zhou, Yifei Liu, Weiming Lu, Xiaozhong Liu, Yating Zhang, Changlong Sun, Fei Wu, Kun Kuang
| Challenge: | Recent advances in deep learning have enabled a variety of techniques to be used to solve the LJP task. |
| Approach: | They propose a framework that leverages the strength of both LLMs and domain-specific models in the context of precedents. |
| Outcome: | The proposed framework leverages the strength of both LLM and domain models in the context of precedents. |
FLRC: Fine-grained Low-Rank Compressor for Efficient LLM Inference (2025.emnlp-main)
Copied to clipboard
| Challenge: | Low-rank compression can reduce memory usage and computational demand, but results are poor during decoding. |
| Approach: | They propose a fine-grained low-rank compression algorithm that determines optimal rank allocation for each layer and incorporates progressive low-ranked decoding to maintain text generation quality. |
| Outcome: | The proposed approach outperforms state-of-the-art methods on summarization tasks and on understanding tasks. |
Beyond One-Size-Fits-All: Tailored Benchmarks for Efficient Evaluation (2025.acl-long)
Copied to clipboard
Peiwen Yuan, Yueqi Zhang, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing efficient methods estimate performance of models on large benchmarks, but these methods rely on the assumption that target models have high prediction consistency with source models. |
| Approach: | They propose a method that conducts customized evaluation tailored to each target model. |
| Outcome: | The proposed method reduces the MAE of estimates by 31.4% on benchmarks across 300 models. |
Beyond Quantity: Trajectory Diversity Scaling for Code Agents (2026.findings-acl)
Copied to clipboard
Guhong Chen, Chenghao Sun, Cheng Fu, Qiyao Wang, Zhihong Huang, ChaoPeng Wei, Guangxu Chen, Feiteng Fang, Ahmadreza Argha, Bing Zhao, Xander Xu, Qi Han, Hamid Alinejad-Rokny, Qiang Qu, Binhua Li, Shiwen Ni, Min Yang, HU Wei, Yongbin Li
| Challenge: | Code large language models (LLMs) are becoming tool-interactive agents . quantity-centric scaling exhibits an early bottleneck that underutilizes trajectory data . et al.: a new approach to scale trajectory diversity improves tool-use generalization . |
| Approach: | They propose a Trajectory Diversity Scaling-based data synthesis framework for code agents that scales performance through diversity rather than raw volume. |
| Outcome: | Experiments on general tool-use benchmarks and code agent tasks show that TDScaling improves tool-user generalization and inherent coding proficiency. |
One for All: Update Parameterized Knowledge Across Multiple Models with Once Edit (2025.acl-long)
Copied to clipboard
Weitao Ma, Xiyuan Du, Xiaocheng Feng, Lei Huang, Yichong Huang, Huiyi Zhang, Xiaoliang Yang, Baohang Li, Xiachong Feng, Ting Liu, Bing Qin
| Challenge: | Existing methods for modifying large language models focus on individual models, resulting in errors and hallucinations. |
| Approach: | They propose an ensemble-based approach that employs a plug-in model as the editing module and a dynamic weight mechanism to enhance its effectiveness. |
| Outcome: | The proposed approach outperforms existing methods while achieving superior editing efficiency. |
TreeReview: A Dynamic Tree of Questions Framework for Deep and Efficient LLM-based Scientific Peer Review (2025.emnlp-main)
Copied to clipboard
Yuan Chang, Ziyue Li, Hengyuan Zhang, Yuanbo Kong, Yanru Wu, Hayden Kwok-Hay So, Zhijiang Guo, Liya Zhu, Ngai Wong
| Challenge: | Large Language Models (LLMs) have shown significant potential in assisting peer review, but current methods struggle to generate thorough and insightful reviews while maintaining efficiency. |
| Approach: | They propose a framework that models paper review as a hierarchical and bidirectional question-answering process. |
| Outcome: | The proposed framework outperforms baselines on full review generation and actionable feedback comments generation tasks while reducing LLM token usage by up to 80% compared to computationally intensive approaches. |
PAD-Net: An Efficient Framework for Dynamic Networks (2023.acl-long)
Copied to clipboard
| Challenge: | Dynamic networks can significantly improve the model’s representation power with acceptable computational cost. |
| Approach: | They propose a partially dynamic network to transform redundant dynamic parameters into static ones and iterative mode partition to partition dynamic and static parameters efficiently. |
| Outcome: | The proposed network surpasses fully dynamic networks by +0.7% top-1 acc with only 30% dynamic parameters for DY-Conv and +1.9% average score in language understanding with only 50% dynamic parameters. |
CoMoL: Efficient Mixture of LoRA Experts via Dynamic Core Space Merging (2026.findings-acl)
Copied to clipboard
Jie Cao, Zhenxuan Fan, Zhuonan Wang, Tianwei Lin, Ziyuan Zhao, Rolan Yan, Wenqiao Zhang, Feifei Shao, Hongwei Wang, Jun Xiao, Siliang Tang
| Challenge: | Existing PEFT methods suffer from limited parameter efficiency and coarse-grained adaptation due to proliferation of LoRA experts and instance-level routing. |
| Approach: | They propose a new MoE-LoRA framework that incorporates expert diversity, parameter efficiency, and fine-grained adaptation. |
| Outcome: | The proposed framework outperforms existing methods on multiple tasks while maintaining parameter efficiency. |
Bridging Local Details and Global Context in Text-Attributed Graphs (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies focus on combining different information levels but overlook interconnections, i.e., contextual textual information among nodes. |
| Approach: | They propose a framework that bridges local and global perspectives by leveraging contextual textual information. |
| Outcome: | The proposed framework achieves state-of-the-art performance while reducing tokens significantly. |
What to Pre-Train on? Efficient Intermediate Task Selection (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for fine-tuning intermediate tasks are inefficient and expensive. |
| Approach: | They propose to use a set of 42 intermediate and 11 target English classification, multiple choice, question answering, and sequence tagging tasks to identify the best settings for intermediate transfer learning. |
| Outcome: | The proposed methods achieve an average Regret@3 of 1% across all target tasks. |
Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing frameworks treat memory as a static append-only archive . Existing systems focus on passive accumulation, resulting in a 'passive accumulation' of memory. |
| Approach: | They propose a framework for experience-driven agent evolution that integrates procedural memory with contextual information to create a high-quality experience pool. |
| Outcome: | Experiments on BFCL-V3 and AppWorld show that ReMe outperforms memoryless Qwen3-8B. |
GTA: Generating Long-horizon Tasks for Web Agents at Scale (2026.acl-long)
Copied to clipboard
Tenghao Huang, Kung-Hsiang Huang, Prafulla Kumar Choubey, Yilun Zhou, Muhao Chen, Jonathan May, Chien-Sheng Wu
| Challenge: | Existing benchmarks provide only coarse start–goal annotations without intermediate trajectories . Existing frameworks provide no supervision over the agent's latent decision process . |
| Approach: | They propose a framework that integrates crawling, retrieval-based seeding, in-context generation and automated quality control to produce realistic tasks paired with executable trajectories. |
| Outcome: | The proposed framework decouples crawling from generation for greater efficiency and ensures dense supervision through deterministic replays and systematic validation. |
From Pseudo-Balancing to True Specialization: Memory-Aware Routing for Mixture-of-Experts (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods to optimize expert-centered load balancing fail to account for pseudo-balance phenomenon . severe knowledge overlap among experts leads to redundant representations and inefficient parameter utilization . |
| Approach: | They propose a method that prioritizes expert utilization over semantic alignment . they use memory-aware routing to ensure expert load balancing is consistent . |
| Outcome: | Experimental results show that MAR improves expert specialization by 35% and accuracy by 2%-25% . MAR matches baseline performance with only half the experts . |
SpecCache: Speculative KV Cache Reuse for Efficient RAG Serving (2026.acl-long)
Copied to clipboard
Zijian Wen, Tao Zhang, Shuangwu Chen, Shenghao Ye, Yu Guo, Qirui Chen, Jingxian Shuai, Yunpeng Hou, Huasen He, null Jianyang
| Challenge: | Retrieval-Augmented Generation (RAG) improves LLMs but faces high prefill latency during long contexts. |
| Approach: | They propose a method that uses deep-layer hidden-state norms to guide token selection . they propose to use deep-layered hidden-status norms as a proxy to guide the token selection. |
| Outcome: | The proposed SpecCache outperforms state-of-the-art (SOTA) benchmarks. |
Human Inspired Progressive Alignment and Comparative Learning for Grounded Word Acquisition (2023.acl-long)
Copied to clipboard
| Challenge: | a recent study shows that word acquisition is an efficient, supervised, and continual process. |
| Approach: | They develop a computational process for word acquisition through comparative learning . they frame the acquisition of words as representation-symbol mapping . |
| Outcome: | The proposed method can be used to learn the meaning of a word efficiently and efficiently. |
Low-Rank Interconnected Adaptation across Layers (2025.findings-acl)
Copied to clipboard
| Challenge: | Low-rank adaptation (LoRA) is a parameter-efficient fine-tuning (PEFT) method that learns weight updates W = AB for pretrained weights W through low-rank adapters A and B. |
| Approach: | They propose a low-rank interconnected adaptation across layers method that introduces an interconnected framework with locally shared A and globally shared B experts. |
| Outcome: | The proposed method improves expressiveness across domains and modalities and enables higher-rank W with equal or fewer parameters. |
Dovetail: A CPU/GPU Heterogeneous Speculative Decoding for LLM inference (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) are demanding more memory and computational resources . however, these devices typically feature weaker GPUs and stronger CPUs . |
| Approach: | They propose a lossless inference acceleration method that leverages the characteristics of heterogeneous devices and the advantages of speculative decoding. |
| Outcome: | The proposed method achieves speedups ranging from 1.79 to 10.1 across different devices . it uses a draft model on the GPU to perform preliminary predictions, while a target model on CPU validates these outputs . |
Multi-Docker-Eval: A ‘Shovel of the Gold Rush’ Benchmark on Automatic Environment Building for Software Engineering (2026.findings-acl)
Copied to clipboard
| Challenge: | Automated environment configuration is a critical bottleneck in scaling software engineering (SWE) automation. |
| Approach: | They propose a reliable evaluation standard for automated environment configuration for 40 real-world repositories spanning 9 programming languages. |
| Outcome: | The proposed benchmark includes 40 real-world repositories spanning 9 programming languages and measures success in achieving executable states and efficiency under realistic constraints. |
Nested-Refinement Metamorphosis: Reflective Evolution for Efficient Optimization of Networking Problems (2025.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) excel in network algorithm design but suffer from inefficient iterative coding and high computational costs. |
| Approach: | They propose a method to iteratively refine task descriptions and metamorphosis on algorithms to generate more effective solutions. |
| Outcome: | Experimental results show that Nested-Refinement Metamorphosis outperforms state-of-the-art approaches in performance and efficiency. |
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts (2023.emnlp-main)
Copied to clipboard
| Challenge: | Extensive experiments show that MEO significantly improves computational efficiency . compared to dense networks, sparsely activated networks only employ a few parameters for each input . |
| Approach: | They propose a method that merges multiple experts into one to reduce computation costs . they demonstrate that a sparse Mixture of Experts (MoE) can reduce the cost by activating a small subset of parameters for each input . |
| Outcome: | The proposed approach reduces the computational cost to that of a single expert by 83.3% compared to 82.6% in vanilla MoE. |
FPE2M2: Approaching Lossless and Efficient Quantization with Native Floating Point (2025.findings-acl)
Copied to clipboard
Ke Yi, Jianwei Zhang, Zhiying Xu, Xinlong Yang, Yang Zhou, Minmin Sun, Zengke Liu, Tong Zhang, Junyang Lin, Jingren Zhou
| Challenge: | Auto-regressive decoding is a memory-bound job, meaning decoding performance is limited by the bandwidth rather than the computational capabilities of the GPU. |
| Approach: | They propose a framework that supports lossless weight-only quantization inference and validate it on Qwen and LLaMA Models. |
| Outcome: | The proposed framework achieves the highest efficiency with lossless accuracy on Qwen and LLaMA Models across various modalities. |
BSFA: Leveraging the Subspace Dichotomy to Accelerate Neural Network Training (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent studies highlight a fundamental dichotomy in deep learning optimization: parameter updates along the top eigendirections of the loss Hessian (Dom-space) capture most of the update magnitude, while updates in the orthogonal component (Bulk-space), have smaller magnitudes but drive most learning progress. |
| Approach: | They propose a plug-and-play framework that scales update components projected onto distinct subspaces and a block-wise strategy that applies this estimation on a per-parameter-block basis. |
| Outcome: | The proposed framework accelerates training by differentially scaling update components projected onto distinct subspaces, while enhancing stability by moderating updates in dominant subspace and boosting convergence speed by amplifying updates in bulk-space. |
Hebbian-Guided Bi-Directional Rank Adaptation for Parameter-Efficient Fine-Tuning (2026.findings-acl)
Copied to clipboard
| Challenge: | Low-Rank Adaptation (LoRA) is a widely used method to fine-tune large language models . but its fixed-rank design cannot capture the varying importance across different layers . |
| Approach: | They propose a framework that bi-directionally reallocates low-rank capacity using Hebbian-inspired importance estimation. |
| Outcome: | Experiments show that HeBiRA improves performance over baselines. |
Distributed LLM Serving on Consumer-Grade GPUs by Reconciling Computation and Communication (2025.findings-emnlp)
Copied to clipboard
Lewei Jin, Kui Zhang, Yongqi Chen, null Zhuoyifan, Renjie Li, Yi Gao, Bowei Yang, Zhengong Cai, Wei Dong
| Challenge: | Large language models are reshaping internet services, and serving them is costly. |
| Approach: | They propose an efficient distributed LLM serving system that splits prefill and decode requests into smaller chunks . |
| Outcome: | The proposed system reduces TTFT, TPOT, and latency compared to the state-of-the-art system. |
IAM: Efficient Inference through Attention Mapping between Different-scale LLMs (2025.acl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) are a challenge due to their internal reasoning processes. |
| Approach: | They propose an algorithm that can optimize attention matrices by performing attention mapping between small and large LLMs. |
| Outcome: | The proposed framework can reduce KV cache usage by 22.1% and accelerate prefill by 15% without sacrificing performance. |
Speculative Streaming: Efficient and Scalable Speculative Decoding with Multi-Stream Attention (2025.emnlp-main)
Copied to clipboard
Nikhil Bhendawade, Irina Belousova, Qichen Fu, Henry Mason, Antonie Lin, Mohammad Rastegari, Mahyar Najibi
| Challenge: | Speculative decoding is a prominent technique for accelerating LLM inference by leveraging an auxiliary draft model, but its effectiveness is limited by the autoregressive nature of draft generation. |
| Approach: | They propose a method that integrates speculative draft generation directly within the target model using multi-stream attention. |
| Outcome: | The proposed method improves acceptance but also latency and speculation latency, limiting overall speedup. |
KVPR: Efficient LLM Inference with I/O-Aware KV Cache Partial Recomputation (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing methods to inference large language models are limited by CPU capabilities and memory constraints. |
| Approach: | They propose an efficient I/O-aware LLM inference method that overlaps GPU computation with KV cache transfer to minimize idle GPU time. |
| Outcome: | The proposed method reduces the cost of auto-regressive decoding by 35.8% . it also achieves 46.2% higher throughput during decoding compared to state-of-the-art methods. |
G-LoRA: Global-Local Decoupled Low-Rank Adaptation (2026.findings-acl)
Copied to clipboard
| Challenge: | Low-Rank Adaptation (LoRA) improves the fine-tuning efficiency and performance of large language models. |
| Approach: | They propose a low-rank adaptive approach that decomposes update matrix into global and local adapters and assigns them to local and global adapters. |
| Outcome: | The proposed method achieves up to 2.7% accuracy improvement over LoRA and its variants on commonsense reasoning, mathematical reasoning, and code generation. |
FLARE: Fine-Grained Length-Aware Routing for Resource-Efficient Heterogeneous LLM Serving (2026.acl-long)
Copied to clipboard
| Challenge: | Existing routers lack fine-grained resource awareness across deployment settings, which degrades efficiency metrics in real-world serving. |
| Approach: | They propose a length-centric, resource-aware multi-LLM routing framework that uses length-based models to estimate per-query latency and cost. |
| Outcome: | Experiments show that FLARE reduces latency and cost by up to 68% and 75% while maintaining competitive accuracy. |
MAESTRO: Meta-learning Adaptive Estimation of Scalarization Trade-offs for Reward Optimization (2026.acl-long)
Copied to clipboard
Yang Zhao, Hepeng Wang, Xiao Ding, Yangou Ouyang, Bibo Cai, Kai Xiong, Jinglong Gao, Zhouhao Sun, Li Du, Bing Qin, Ting Liu
| Challenge: | Group-Relative Policy Optimization (GRPO) has emerged as an efficient paradigm for aligning Large Language Models (LLMs), but its efficacy is confined to domains with verifiable ground truths. |
| Approach: | They propose a meta-cognitive orchestration layer that treats reward scalarization as a dynamic latent policy, leveraging the model’s terminal hidden states as 'a semantic bottleneck' . Across seven benchmarks, MAESTRO consistently outperforms single-reward and static multi-objective baselines while preserving the efficiency advantages of GRPO. |
| Outcome: | The proposed model outperforms single-reward and static multi-objective baselines while preserving efficiency advantages. |
Bit-by-Bit: Progressive QAT Strategy with Outlier Channel Splitting for Stable Low-Bit LLMs (2026.acl-long)
Copied to clipboard
Binxing Xu, Hao Gu, Lujun Li, Hao Wang, Bei Liu, Jiacheng Liu, Qiyuan Zhu, Xintong Yang, Chao Li, Sirui Han, Yike Guo
| Challenge: | Existing approaches to training LLMs at ultra-low precisions suffer from convergence instability and substantial training costs. |
| Approach: | They propose a progressive QAT framework with outlier channel splitting to address these issues . they use nested structure of integer quantization grids to enable a "train once, deploy any precision" paradigm . |
| Outcome: | The proposed framework outperforms baselines on both Llama2/3 and W2A16, with an 11 speedup over BF16. |
Composable Text Controls in Latent Space with ODEs (2023.emnlp-main)
Copied to clipboard
Guangyi Liu, Zeyu Feng, Yuan Gao, Zichao Yang, Xiaodan Liang, Junwei Bao, Xiaodong He, Shuguang Cui, Zhen Li, Zhiting Hu
| Challenge: | Existing approaches to composable text operations often require plug-and-play . a single LM can perform arbitrary text operation composition in the latent space . |
| Approach: | They propose an efficient approach for composable text operations in the latent space of text . they connect pretrained LMs to the laten space and adapt them to the space . |
| Outcome: | The proposed approach improves on existing methods in the latent space of text. |
SPIDE: Serial and Parallel Intertwined Speculative Decoding (2026.findings-acl)
Copied to clipboard
Wenru Xu, Peixuan Xu, Ziqi Yang, Ming Hu, Zihui Wang, Jianzhong Qi, Rongshan Yu, Xiaoliang Fan, Cheng Wang
| Challenge: | Speculative decoding (SD) is a training-free SD framework that orchestrates dynamic alternation combining serial dynamic drafting with parallel draft verification. |
| Approach: | They propose a serial and parallel intertwined speculative DEcoding framework that orchestrates dynamic alternation combining serial dynamic drafting and parallel draft verification. |
| Outcome: | The proposed framework accelerates inference while reducing the LLM usage costs. |
World Modeling Makes a Better Planner: Dual Preference Optimization for Embodied Task Planning (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches focus on action selection or use pre-trained models as world models to enhance planning capabilities. |
| Approach: | They propose a new learning framework that optimizes state prediction and action selection through preference learning. |
| Outcome: | The proposed method outperforms existing methods and GPT-4o on VoTa-Bench and Qwen2-VL (7B), LLaVA-1.6 (7B) and LLama-3.2 (11B). |
Adaptive Preference Optimization with Uncertainty-aware Utility Anchor (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Offline preference optimization methods are efficient for large language models (LLMs) alignment. |
| Approach: | They propose an offline preference optimization framework that estimates uncertainties from preference data . the method enables training even in scenarios where the data is unpaired . |
| Outcome: | The proposed method enables training even in scenarios where the data is unpaired . |
TENP: Trapezoidal Expert Neuron Pruning For Mixture-of-Experts (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing compression approaches remove entire experts, disrupting routing topology and harming performance, or rely on unstructured weight pruning with limited practical efficiency. |
| Approach: | They propose a structured **T**rapezoidal **E**xpert **N**euron **P**running framework that uses a trapezoidal pattern to identify and retain important experts while applying expert neuron pruning (ENP) to less important experts. |
| Outcome: | The proposed framework outperforms the full-parameter model by 10% on code generation tasks under a sparse activation of experts and a 40% routing sparsity. |
NormAL LoRA: What is the perfect size? (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are crucial for enabling intelligent experiences across applications. |
| Approach: | They propose a low-rank adaptive localization method that uses rank-norm regularization to determine the optimal rank for each weight matrix. |
| Outcome: | NormAL LoRA reduces adapter parameters by 37% while preserving full fine-tuning performance. |
Comprehensive and Efficient Distillation for Lightweight Sentiment Analysis Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent efforts to develop lightweight and practical sentiment analysis models are limited by manual instruction and large-scale user texts. |
| Approach: | They propose a framework for sentiment analysis that uses attribute-based instruction construction and difficulty-based data filtering to distill knowledge. |
| Outcome: | The proposed framework outperforms baseline methods in data efficiency and performance. |
DRAE: Dynamic Retrieval-Augmented Expert Networks for Lifelong Learning and Task Adaptation in Robotics (2025.acl-long)
Copied to clipboard
| Challenge: | Experimental results show that Dynamic Retrieval-Augmented Expert Networks outperforms baseline approaches in long-term task retention and knowledge reuse. |
| Approach: | They propose a dynamic routing architecture that leverages MoE and Retrieval-Augmented Generation to augment the learning process. |
| Outcome: | The proposed architecture outperforms baseline approaches in long-term task retention and knowledge reuse. |
Targeted Distillation for Sentiment Analysis (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent studies demonstrate that large language models exhibit remarkable capabilities and achieve state-of-the-art performance in diverse sentiment analysis tasks. |
| Approach: | They propose a distillation framework that decouples knowledge from alignment and introduces a sentiment analysis benchmark that covers a diverse set of tasks. |
| Outcome: | The proposed framework improves models' generalization to unseen tasks and their generalization is strong against existing small-scale models. |
Inefficiencies of Meta Agents for Agent Design (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Recent work has automated the design of agentic systems using meta-agents . authors examine three key challenges in a common class of meta-gents. |
| Approach: | They examine how meta-agents learn across iterations and show performance improves with evolutionary approach. |
| Outcome: | The proposed meta-agents perform worse when iterating on multiple agents than human-designed agents. |
ChipSeek: Optimizing Verilog Generation via EDA-Integrated Reinforcement Learning (2026.acl-long)
Copied to clipboard
Zhirong Chen, Kaiyan Chang, Zhuolin Li, Cangyuan Li, Xinyang He, Chujie Chen, Mengdi Wang, Haobo Xu, Yinhe Han, Huawei Li, Ying Wang
| Challenge: | Existing approaches to optimize Register-Transfer Level (RTL) code fail to simultaneously optimize functional correctness and hardware efficiency metrics such as Power, Performance, and Area (PPA). |
| Approach: | They propose a hierarchical reward based reinforcement learning framework that integrates direct feedback from EDA simulators and synthesis tools into a reward mechanism. |
| Outcome: | The proposed framework integrates direct feedback from EDA simulators and synthesis tools into a hierarchical reward based reinforcement learning framework. |
TASO: Task-Aligned Sparse Optimization for Parameter-Efficient Model Adaptation (2025.emnlp-main)
Copied to clipboard
Daiye Miao, Yufang Liu, Jie Wang, Changzhi Sun, Yunke Zhang, Demei Yan, Shaokang Dong, Qi Zhang, Yuanbin Wu
| Challenge: | Existing studies have shown that LoRA introduces substantial parameter redundancy, which not only increases the number of trainable parameters but also hinders the effectiveness of fine-tuning. |
| Approach: | They propose a method that leverages importance information from the pretrained model’s weights to mitigate LoRA redundancy. |
| Outcome: | The proposed method significantly reduces the number of trainable parameters required for task adaptation while providing a task-aligned perspective for LoRA redundancy reduction. |
Mixture of Weight-shared Heterogeneous Group Attention Experts for Dynamic Token-wise KV Optimization (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for token-level KV optimization and grouping of tokens are inefficient and strain compute and storage resources. |
| Approach: | They propose a mixture-of-expert approach that dynamically optimizes token-wise computation and memory allocation by a token-based expert-choice routing mechanism guided by learned importance scores. |
| Outcome: | The proposed approach retains all tokens while adaptively routing them to specialized experts with varying KV group sizes, balancing granularity and efficiency. |
Don’t Get Lost in the Trees: Streamlining LLM Reasoning by Overcoming Tree Search Exploration Pitfalls (2025.acl-long)
Copied to clipboard
Ante Wang, Linfeng Song, Ye Tian, Dian Yu, Haitao Mi, Xiangyu Duan, Zhaopeng Tu, Jinsong Su, Dong Yu
| Challenge: | Recent advances in tree search algorithms guided by verifiers have significantly enhanced the reasoning capabilities of large language models (LLMs), but at the cost of increased computational resources. |
| Approach: | They propose an e ffici ent tree sear ch framework that is a plug-and-play system compatible with various tree search algorithms. |
| Outcome: | The proposed framework reduces computational costs and prioritizes resource allocation to harder tasks (Levels 3-4) over simpler ones (Level 1-2), addressing both over-exploration in basic problems and under-exploation in complex cases. |
Memorize Step by Step: Efficient Long-Context Prefilling with Incremental Memory and Decremental Chunk (2024.emnlp-main)
Copied to clipboard
Zhiyuan Zeng, Qipeng Guo, Xiaoran Liu, Zhangyue Yin, Wentao Shu, Mianqiu Huang, Bo Wang, Yunhua Zhou, Linlin Li, Qun Liu, Xipeng Qiu
| Challenge: | Existing methods to optimize LLM for long sequences for long documents are slow and consume memory. |
| Approach: | They propose a method that starts with a small memory size and gradually increases it . they propose Decremental Chunk based on Incremental Memory (IMDC) which reduces chunk size while increasing memory size . |
| Outcome: | The proposed method is faster (1.45x) and reduces GPU memory consumption by 23.3% compared to fixed-size memory. |
GRaMPa: Subword Regularisation by Skewing Uniform Segmentation Distributions with an Efficient Path-counting Markov Model (2025.acl-long)
Copied to clipboard
| Challenge: | Subword regularisations are known to be stochastic, but only a handful of possible segmentations are sampled. |
| Approach: | They propose to randomise word segmentations from a subword tokeniser instead of randomising them by weighting paths in an unweighted segmentation graph. |
| Outcome: | The proposed method outperforms existing methods on token-level tasks with spelling errors. |
ElasticFlow: One-Step Physics-Consistent Policy with Elastic Time Horizons for Language-Guided Manipulation (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for embodied AI use iterative denoising to achieve high latency and lack physical consistency. |
| Approach: | They propose a distillation-free, physics-consistent one-step policy framework that reconstructs the Mean Field Theory by directly modeling the average velocity field. |
| Outcome: | Experiments on LIBERO, CALVIN, and RoboTwin show that the proposed framework outperforms state-of-the-art methods on long-horizon tasks. |
Localized Low-Rank Adaptation within Clustered Parameter Subspaces (2026.acl-long)
Copied to clipboard
| Challenge: | Low-Rank Adaptation (LoRA) for large language models has been successful in various domains. |
| Approach: | They propose to perform low-rank updates within clustered parameter subspaces . they group rows/columns of update matrix into locally coherent, uncorrelated subspace blocks . |
| Outcome: | Empirical results show that low-rank Adaptation (LoRA) is better than global adaptations in various domains. |
Stability Implies Redundancy: Delta Attention Selective Halting for Efficient Long-Context Prefilling (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods to reduce sequence length rely on heuristics that break compatibility with hardware-efficient kernels like FlashAttention. |
| Approach: | They propose a method that selectively halts stabilized tokens by monitoring layer-wise update dynamics of the self-attention mechanism. |
| Outcome: | The proposed method can reduce prefill complexity while preserving model accuracy and hardware efficiency. |
A Data-Efficient Path to Multilingual LLMs: Language Expansion via Post-training PARAM𝛥 Integration into Upcycled MoE (2026.acl-long)
Copied to clipboard
Hao Zhou, Tianhao Li, Zhijun Wang, Shuaijie She, Linjuan Wu, Hao-Ran Wei, Baosong Yang, Jiajun Chen, Shujian Huang
| Challenge: | Large Language Models (LLMs) are expensive and require extensive Continued Pre-Training and data-intensive alignment to expand. |
| Approach: | They propose a method which upcycles a dense model into a Mixture-of-Experts architecture, allocating different experts to different languages. |
| Outcome: | Experiments show that the proposed model upcycles a dense model into a Mixture-of-Experts(MoE) architecture, allocating different experts to different languages. |
HATA: Trainable and Hardware-Efficient Hash-Aware Top-k Attention for Scalable Large Model Inference (2025.findings-acl)
Copied to clipboard
Ping Gong, Jiawei Yi, Shengnan Wang, Juncheng Zhang, Zewen Jin, Ouxiang Zhou, Ruibo Liu, Guanbin Xu, Youhui Bai, Bowen Ye, Kun Yuan, Tong Yang, Gong Zhang, Renhai Chen, Feng Wu, Cheng Li
| Challenge: | Existing top-k attention methods struggle to strike a balance between efficiency and accuracy. |
| Approach: | They propose a top-k attention approach that integrates low-overhead techniques into the Top-k Attention process to achieve 7.2 speedup compared to vanilla full attention. |
| Outcome: | The proposed approach achieves 7.2 speedup compared to current top-k attention methods while maintaining model accuracy. |
The Efficiency vs. Accuracy Trade-off: Optimizing RAG-Enhanced LLM Recommender Systems Using Multi-Head Early Exit (2025.acl-long)
Copied to clipboard
Huixue Zhou, Hengrui Gu, Zaifu Zhan, Xi Liu, Kaixiong Zhou, Yongkang Xiao, Mingfu Liang, Srinivas Prasad Govindan, Piyush Chawla, Jiyan Yang, Xiangfei Meng, Huayu Li, Buyun Zhang, Liang Luo, Wen-Yen Chen, Yiping Han, Bo Long, Rui Zhang, Tianlong Chen
| Challenge: | Existing frameworks for Large Language Models (LLMs) for Click-Through Rate prediction require a careful balance between computational efficiency and predictive accuracy. |
| Approach: | They propose a framework that integrates Retrieval-Augmented Generation with a novel multi-head early exit architecture to address both challenges. |
| Outcome: | The proposed framework reduces retrieval time while maintaining high model performance. |
AgentAsk: Multi-Agent Systems Need to Ask (2026.acl-long)
Copied to clipboard
Bohan Lin, Kuo Yang, Zelin Tan, Yingchuan Lai, Chen Zhang, Guibin Zhang, Xinlei Yu, Miao Yu, Xu Wang, Yudong Zhang, Yang Wang
| Challenge: | Multi-agent systems fail to consistently outperform strong single-a agent baselines due to error propagation at inter-aggent message handoffs. |
| Approach: | They propose an edge-level error taxonomy that identifies four main errors in multi-agent interactions as data gaps, signal corruption, referential drift and capacity gaps as primary sources of failure. |
| Outcome: | The proposed module outperforms existing systems on five benchmarks and is architecture-agnostic. |
Enhancing RAG Efficiency with Adaptive Context Compression (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods apply fixed compression rates, over-compressing simple queries or under-compressed complex ones. |
| Approach: | a new framework uses a hierarchical compressor and a context selector to optimize inference efficiency . a framework that dynamically adjusts compression rates based on input complexity optimizes inference without loss of accuracy. |
| Outcome: | Adaptive Context Compression for RAG outperforms fixed-rate methods on Wikipedia and five QA datasets . |
Bone Soups: A Seek-and-Soup Model Merging Approach for Controllable Multi-Objective Generation (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches focus on merging language models tuned on single objectives . existing approaches ignore the impacts of competing objectives on model tuning . |
| Approach: | They propose a model merging approach that seeks a series of backbone models and merges them according to user preferences. |
| Outcome: | The proposed approach exhibits strong controllability and Pareto optimality in controllable multi-objective generation. |
Nexus: Adaptive Upcycling to Efficiently Pretrain Mixture of Experts (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Nevertheless, training from scratch on trillions of tokens remains expensive that most users can only finetune these models. |
| Approach: | They propose to reuse parameters of dense models for the MoE layers with a router that can integrate new experts into existing trained models without hurting performance on previous domains. |
| Outcome: | The proposed router can integrate new experts into existing trained models without hurting the performance on previous domains. |
Learn Like Humans: Use Meta-cognitive Reflection for Efficient Self-Improvement (2026.acl-long)
Copied to clipboard
| Challenge: | Existing self-improving frameworks rely on inefficient, multi-turn recursive loops that incur high computational costs. |
| Approach: | They propose a framework that achieves efficient self-evolution within a single recurrence cycle. |
| Outcome: | The proposed framework outperforms state-of-the-art self-evolving systems while significantly reducing computational overhead. |
MMDEND: Dendrite-Inspired Multi-Branch Multi-Compartment Parallel Spiking Neuron for Sequence Modeling (2025.acl-long)
Copied to clipboard
Kexin Wang, Yuhong Chou, Di Shang, Shijie Mei, Jiahong Zhang, Yanbin Huang, Man Yao, Bo Xu, Guoqi Li
| Challenge: | Vanilla spiking neurons are simplified from complex biological neurons with dendrites, soma, and synapses into single somatic compartments. |
| Approach: | They propose a multi-branch, multi-compartment parallel spiking dendritic neuron with a proportion-adjustable multi-branched structure that enables long-term temporal dependencies. |
| Outcome: | The proposed model achieves better long-sequence modeling capability with fewer parameters and lower energy consumption. |
Stabilizing Efficient Reasoning with Step-Level Advantage Selection (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models generate long and verbose reasoning traces at inference time . short context post-training alone induces substantial reasoning compression . |
| Approach: | They propose a step-level advantage selection approach that reduces reasoning length by over 30% . they propose to use GRPO without any length-aware objective to train models in a shorter context window . |
| Outcome: | The proposed approach reduces average reasoning length by over 30% while improving Pass@1 accuracy by 3.79 points over the strongest length-aware baseline. |
EGSS: Entropy-guided Stepwise Scaling for Reliable Software Engineering (2026.acl-long)
Copied to clipboard
Chenhui Mao, Yuanting Lei, Zhixiang Wei, Ming Liang, Zhixiang Wang, Jingxuan Xu, Dajun Chen, Wei Jiang, Yong Li
| Challenge: | Entropy-Guided Stepwise Scaling (EGSS) is a novel TTS framework for software engineering tasks. |
| Approach: | They propose an entropy-guided stepwise scaling framework that balances efficiency and effectiveness through entropic-guide encoding and robust test-suite augmentation. |
| Outcome: | EGSS boosts performance by 5–10% across all evaluated models, and reduces inference-time token usage by over 28% . compared to existing methods, EGS reduces token usage and reduce inference time by over 20% . |
HydraOpt: Navigating the Efficiency-Performance Trade-off of Adapter Merging (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods that produce a fixed trade-off between storage size and performance are often ineffective due to the growing size of large language models. |
| Approach: | They propose a model merging technique that capitalizes on similarities between low-rank adapters to reduce storage costs and improve performance. |
| Outcome: | The proposed method significantly reduces storage size (48% reduction) while outperforms existing merging techniques in terms of performance (0.2-1.8% drop). |
Calibrated Speculative Decoding: Frequency-Guided Candidate Selection for Efficient Inference (2026.acl-long)
Copied to clipboard
| Challenge: | Speculative decoding (SD) is a powerful and efficient way to accelerate autoregressive generation. |
| Approach: | They propose a training-free framework that recovers valid tokens discarded by standard verification . they use online correction memory and Semantic Consistency Gating to analyze rejections . |
| Outcome: | The proposed framework outperforms existing methods and achieves peak throughput speedup of 2.33x. |
ParaCook: On Time-Efficient Planning for Multi-Agent Systems (2026.findings-acl)
Copied to clipboard
Shiqi Zhang, Xinbei Ma, Yunqing Xu, Zouying Cao, Pengrui Lu, Haobo Yuan, Tiancheng Shen, Zhuosheng Zhang, Hai Zhao, Ming-Hsuan Yang
| Challenge: | Existing agent benchmarks focus on task completion while neglecting time efficiency in parallel and asynchronous operations. |
| Approach: | They propose a framework for large language models that allows agents to plan long-horizon tasks in a scalable way. |
| Outcome: | The proposed framework is based on the Overcooked game and can be used to evaluate time efficiency-aware multi-agent planning. |
Automated Fine-Grained Mixture-of-Experts Quantization (2025.findings-acl)
Copied to clipboard
| Challenge: | specialized quantization framework for Mixture of Experts architectures is inadequate for model compression. |
| Approach: | They propose a specialized quantization framework for Mixture of Experts architectures . they find that expert networks exhibit distinctive channel-wise outlier distributions ." |
| Outcome: | The proposed framework improves on the Mixtral-8x7b-v0.1 architecture while maintaining minimal computational overhead. |
Improving Neutral Point-of-View Generation with Data- and Parameter-Efficient RL (2025.emnlp-main)
Copied to clipboard
Jessica Hoffmann, Christiane Ahlheim, Zac Yu, Aria Walfrand, Jarvis Jin, Marie Tano, Ahmad Beirami, Erin MacMurray van Liemt, Nithum Thain, Hakim Sidahmed, Lucas Dixon
| Challenge: | Parameter-efficient reinforcement learning (PE-RL) is a highly effective training regime to improve large language models’ ability to answer queries on sensitive topics with a Neutral Point of View (NPOV). |
| Approach: | They propose to use parameter-efficient reinforcement learning to train large language models to answer queries with a Neutral Point of View (NPOV) This is compared to the strongest baseline, LoRA finetuning, SFT and RLHF. |
| Outcome: | The proposed training regime improves on NPOV quality and scores higher on features identified by linguists as key to separating good answers from the best answers. |
End-to-End Optimization of LLM-Driven Multi-Agent Search Systems via Heterogeneous-Group-Based Reinforcement Learning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing multi-agent reinforcement learning methods depend on large critic networks to evaluate joint actions, leading to instability and high memory costs. |
| Approach: | They propose a method to optimize large language models for agent-specific roles . they propose combining agent-based frameworks with retrieval-augmented generation . |
| Outcome: | Experiments show that multi-agent group policy optimization outperforms baselines in task performance and computational efficiency. |
AutoSearch: Adaptive Search Depth for Efficient Agentic RAG via Reinforcement Learning (2026.findings-acl)
Copied to clipboard
Jingbo Sun, Wenyue Chong, Songjun Tu, Qichao Zhang, Yaocheng Zhang, Jiajun Chai, Xiaohan Wang, Wei Lin, Guojun Yin, Dongbin Zhao
| Challenge: | Prior work limits search depth to reduce cost, but this often leads to underexploration of complex questions. |
| Approach: | They propose a reinforcement learning framework that evaluates each search step via self-generated intermediate answers. |
| Outcome: | Extensive experiments on multiple benchmarks show that AutoSearch achieves a superior accuracy-efficiency trade-off, alleviating over-searching while preserving search quality. |
From Outcome to Process: Optimizing MoE Load Balancing with MCTS (2026.findings-acl)
Copied to clipboard
Wenjun Ke, Hengyuan Xu, Ziyu Shang, Yao He, Jiahao Wang, Zijie Xu, Peng Wang, Yuhang Lou, Jiajun Liu
| Challenge: | Existing balancing strategies focus on constraining the final distribution of expert usage, but overlook the routing decisions made at each layer. |
| Approach: | They propose a three-stage framework that leverages process-level rewards to guide balanced expert routing. |
| Outcome: | Extensive experiments show that LayerMoE improves the performance of state-of-the-art LoRA-MoA baselines, yielding an average accuracy gain of 1.39%. |
SubLIME: Subset Selection via Rank Correlation Prediction for Data-Efficient LLM Evaluation (2025.acl-long)
Copied to clipboard
Gayathri Saranathan, Cong Xu, Mahammad Parwez Alam, Tarun Kumar, Martin Foltin, Soon Yee Wong, Suparna Bhattacharya
| Challenge: | Large language models and datasets have made benchmark evaluations computationally prohibitive. |
| Approach: | They propose a framework that reduces evaluation costs by 80% to 99% while preserving ranking fidelity. |
| Outcome: | The proposed evaluation reduces evaluation costs by 80% to 99% while preserving ranking fidelity. |
TACLR: A Scalable and Efficient Retrieval-based Method for Industrial Product Attribute Value Identification (2025.acl-long)
Copied to clipboard
Yindu Su, Huike Zou, Lin Sun, Ting Zhang, Haiyang Yang, Chen Li Yu, David Lo, Qingheng Zhang, Shuguang Han, Jufeng Chen
| Challenge: | Existing methods for product attribute value identification face critical challenges . seller-provided attribute values are often incomplete or inaccurate . |
| Approach: | They propose a retrieval-based method that uses taxonomy-aware contrastive learning . they use product profiles and candidate values to encode and retrieve attributes based on similarity . |
| Outcome: | The proposed method is based on a taxonomy-aware, hard negative sampling and adaptive inference with dynamic thresholds. |
STEP: Success-Rate-Aware Trajectory-Efficient Policy Optimization (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing GRPO-based methods allocate sampling uniformly across tasks regardless of difficulty, propagate misleading learning signals and incur high sample-collection costs. |
| Approach: | They propose a framework that allocates sampling based on per-task success rates and performs fine-grained step-level optimization. |
| Outcome: | The proposed method improves sample efficiency and training stability over existing GRPO variants and three ablation variants on OSWorld and AndroidWorld. |
LeanK: Learnable K Cache Channel Pruning for Efficient Decoding (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing efforts to optimize the key-value (KV) cache include: (1) Eviction, which discards cache of less important tokens; (2) Selection, which retains the full KV cache but selectively reads relevant entries. |
| Approach: | They propose a learning-based method that prunes unimportant key (K) cache channels by leveraging static channel sparsity. |
| Outcome: | Experiments show that LeanK reduces GPU memory and accelerates decoding without sacrificing accuracy. |
FastKV: Decoupling of Context Reduction and KV Cache Compression for Prefill-Decoding Acceleration (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) excel at handling long-context sequences, but require substantial prefill computation and key-value (KV) cache. |
| Approach: | They propose a KV cache compression framework that decouples prefill computation from decoding KV budget. |
| Outcome: | The proposed framework reduces latency in prefill and decoding by leveraging the stabilization of token importance in later layers. |
RecMem: Recurrence-based Memory Consolidation for Efficient and Effective Long-Running LLM Agents (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing memory systems invoke LLMs to extract episodic and semantic memory, and this leads to substantial token consumption. |
| Approach: | They propose a method that stores incoming interactions in a subconscious memory layer and encodes them using lightweight embedding models for retrieval. |
| Outcome: | Experiments show that RecMem reduces the memory construction token cost of three SOTA memory systems by up to 87% while exceeding their accuracy. |
Orthogonal Finetuning Made Scalable (2025.emnlp-main)
Copied to clipboard
| Challenge: | a recent shift in foundation models has slowed the adoption of finetuning methods . however, its high runtime and memory demands limit its scalability . |
| Approach: | They propose an input-centric reformulation that uses matrix-vector multiplications instead of cubic multiplication . they extend OFTv2 to support finetuning quantized foundation models and show it outperforms QLoRA . |
| Outcome: | The proposed model outperforms the popular QLoRA in training stability, efficiency, and memory usage. |
Attn-GS: Attention-Guided Context Compression for Efficient Personalized LLMs (2026.acl-long)
Copied to clipboard
Shenglai Zeng, Tianqi Zheng, Chuan Tian, Dante Everaert, Yau-Shian Wang, Yupin Huang, Michael J. Morais, Rohit Patki, Jinjin Tian, Xinnan Dai, Kai Guo, Monica Xiao Cheng, Hui Liu
| Challenge: | Existing approaches to personalize large language models (LLMs) rely on heuristic methods to compress user profiles but they ignore how LLMs process and prioritize different profile components. |
| Approach: | They propose an attention-guided context compression framework that leverages attention feedback from a marking model to mark important personalization sentences and guides a compression model to generate task-relevant compressed user contexts. |
| Outcome: | The proposed framework outperforms baselines across tasks, token limits, and settings while reducing token usage by 50 times. |
Union-of-Experts: Neurons in Mixture-of-Experts are Secretly Routers (2026.acl-long)
Copied to clipboard
| Challenge: | Mixture-of-Experts (MoE) models rely on an external router to assign tokens to experts, resulting in suboptimal performance. |
| Approach: | They propose an MoE variant that performs "expert-autonomous routing" by pre-designating a fraction of neurons within each expert as "routing neurons" they pre-train UoE models with up to 3B parameters and show they outperform traditional MoEs with matched efficiency. |
| Outcome: | The proposed model outperforms existing models with 3B parameters and provides valuable insights into expert-autonomous selection and the broader routing mechanisms of MoE models. |
Dynamic Tool Dependency Retrieval for Lightweight Function Calling (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing retrieval methods rely on static inputs, failing to capture multi-step tool dependencies and evolving task context. |
| Approach: | They propose a lightweight retrieval method that conditions on initial query and evolving task context. |
| Outcome: | The proposed method improves function calling success rates between 23% and 104% compared to state-of-the-art retrieval methods. |
GUI-Bee: Align GUI Action Grounding to Novel Environments via Autonomous Exploration (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent work of GUI action grounding fine-tunes data from pre-trained MLLMs, but data is limited to specific GUI environments. |
| Approach: | They propose to use a GUI-based agent to collect environment-specific data and fine-tune GUI grounding models with the collected data. |
| Outcome: | The proposed model can be extended to other GUI environments to improve performance. |
DPLoRA: A Dual-Pruning Framework based on ILP Optimization and Progressive Pruning for Parameter-Efficient LoRA Fine-Tuning (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) require computational resources for fine-tuning. |
| Approach: | They propose a framework that optimizes rank allocation via two stages . they propose an initial pruning stage and a progressive pruning stage . |
| Outcome: | The proposed framework outperforms existing PEFT baselines on GLUE and instruction-following tasks while reducing training time and trainable parameters by over 80%. |
AOT*: Efficient Synthesis Planning via LLM-Empowered AND-OR Tree Search (2026.findings-acl)
Copied to clipboard
| Challenge: | Experimental evaluation shows that AOT* achieves competitive solve rates using 3-5 fewer iterations than existing LLM-based approaches. |
| Approach: | They propose a framework that integrates LLM-generated chemical synthesis pathways with systematic AND-OR tree search. |
| Outcome: | Experimental results show that AOT* improves search efficiency and solves faster than existing approaches. |
Dynamic Generation of Multi LLM Agents Communication Topologies with Graph Diffusion Models (2026.acl-long)
Copied to clipboard
Eric Hanchen Jiang, Levina Li, Frank Wan, Xiao Liang, Sophia Yin, Yuchen Wu, Xinfeng Li, Yizhou Sun, Wei Wang, Kai-Wei Chang, Ying Nian Wu
| Challenge: | Existing frameworks rely on static or rule-based topologies that fail to adapt to task requirements. |
| Approach: | They propose a generative framework that generates highly task-adaptive topologies . they validated the framework on multiple benchmarks and validated it on multiple platforms . |
| Outcome: | The proposed framework outperforms existing frameworks in task-adaptive communication topologies. |
FLOP-Efficient Training: Early Stopping Based on Test-Time Compute Awareness (2026.findings-acl)
Copied to clipboard
| Challenge: | Prior work shows that increasing test-time compute (TTC) can improve accuracy of large language models. |
| Approach: | They propose a TTC-aware training algorithm that jointly selects a checkpoint and a corresponding TTC configuration to minimize training compute without sacrificing accuracy. |
| Outcome: | The proposed method reduces training compute by 92% while maintaining accuracy. |
Efficient and Effective Internal Memory Retrieval for LLM-Based Healthcare Prediction (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to ground large language models in external knowledge are limited by hallucinations and a lack of granular medical context. |
| Approach: | They propose a framework that replaces external retrieval with internal, key-based knowledge access by encoding clinical information directly into the model’s parameter space. |
| Outcome: | The proposed framework achieves state-of-the-art performance across four benchmark healthcare outcome prediction datasets. |
Speculative Verification: Exploiting Information Gain for Speculative Decoding (2026.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are used for many applications but their size and computational cost make inference serving a significant challenge. |
| Approach: | They propose an efficient augmentation to Speculative Decoding (SD) that predicts speculation accuracy and dynamically adapts the verification length to maximize throughput. |
| Outcome: | The proposed model reduces wasted verification on rejected tokens and improves decoding efficiency. |
PDTrim: Targeted Pruning for Prefill-Decode Disaggregation in Inference (2026.acl-long)
Copied to clipboard
| Challenge: | Existing pruning methods ignore prefill-decode (PD) disaggregation in practice. |
| Approach: | They propose a pruning method that is highly integrated with prefill-decode (PD) disaggregation, enabling more precise pruning of blocks. |
| Outcome: | The proposed method achieves strong performance in both PD disaggregation and PD unified settings, and can be extended to other non-block pruning methods. |
AdaMix: Adaptive Mixing for Short and Long Reasoning Adapters (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for large reasoning models have improved efficiency but still face limitations such as conflicting objectives and limited adaptability. |
| Approach: | They propose an adaptive reasoning framework that applies a uniform, computation-intensive deep reasoning strategy to all problems. |
| Outcome: | The proposed framework reduces the average response length of DeepSeek-R1-Distill-Qwen-7B by 54.9% while improving accuracy by up to 4.8% on five mathematical datasets. |
Distilling Long-CoT Reasoning through Collaborative Step-wise Multi-Teacher Decoding (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing curation-based approaches to inference are inefficient and fail to adapt dynamically, leading to redundant sampling and missed opportunities for complementary reasoning. |
| Approach: | They propose a collaborative multi-teacher decoding framework that performs step-wise reasoning synthesis guided by predictive perplexity–based scoring and beam search. |
| Outcome: | The proposed framework generates higher-quality reasoning data and achieves student-level results, demonstrating that fine-grained collaboration yields structured, efficient, and robust reasoning distillation. |
Beyond Compromise: Pareto-Lenient Consensus for Efficient Multi-Preference LLM Alignment (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent approaches to align LLMs with diverse human values are based on static linear scalarization or rigid gradient projection . however, these approaches often sacrifice potential global Pareto improvements to avoid transient local trade-offs. |
| Approach: | They propose a game-theoretic framework that reimagines alignment as a dynamic negotiation process. |
| Outcome: | The proposed framework breaks the deadlock between static linear scalarization and rigid gradient projection . it allows the model to escape local degradation and explore the distal Pareto-optimal frontier . |
Thinking Economically: A Hierarchical Framework for Adaptive-Complexity Reasoning in LLMs (2026.findings-acl)
Copied to clipboard
Yubo Gao, Haotian Wu, Hong Chen, Junquan Huang, Yibo Yan, Jungang Li, Zihao Dongfang, Sicheng Tao, PS Tan, Jie Zhang, Xuming Hu
| Challenge: | Existing efficiency methods for Chain-of-Thought (CoT) generate excessively long rationales without commensurate accuracy gains. |
| Approach: | They propose a training framework that operationalizes this principle through coarse-to-fine budgeting. |
| Outcome: | Experiments on GSM8K and MATH500 show that HAB surpasses standard CoT in accuracy and reduces token usage, achieving stronger performance-efficiency trade-off than baselines. |
ProMCP: Profiling Token Flows and Latency Costs in Model Context Protocol–Based LLM Agents (2026.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models are increasingly used as agents that interact with external tools and data sources to solve tasks that require fresh knowledge, precise computation, or action in a real environment. |
| Approach: | They propose a framework that decomposes the MCP workflow into a six-stage communication pipeline and enables granular attribution of computational costs. |
| Outcome: | The proposed framework decomposes the MCP workflow into a six-stage communication pipeline, enabling granular attribution of computational costs. |
Skill Weaving: Efficient LLM Improvement via Modular Skillpacks (2026.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) can specialize under fixed memory and inference budgets, but they struggle to achieve high performance across heterogeneous domains. |
| Approach: | They propose a modular improvement framework that partitions full capabilities of a general-purpose model into domain-specific delta modules that reorganize and refine the model's internal knowledge. |
| Outcome: | The proposed framework outperforms monolithic models on multi-task and agentic benchmarks and achieves up to 4 speedup. |
LipoAgent: Coordinating Fine-Tuned LLM Agents for Safer Lipid Design (2026.findings-acl)
Copied to clipboard
| Challenge: | Lipid nanoparticles (LNPs) are among the most clinically mature platforms for nucleic acid delivery, yet designing lipids that are effective and biologically safe remains a major bottleneck. |
| Approach: | They propose a safety-aware multi-agent LLM framework for lipid discovery that enforces toxicity as a prerequisite for efficiency prediction. |
| Outcome: | The proposed framework achieves an average improvement in mRNA transfection efficiency prediction across multiple foundation models. |
Universally Empowering Zeroth-Order Optimization via Adaptive Layer-wise Sampling (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for fine-tuning Large Language Models are slow and lack of performance. |
| Approach: | They propose a Zeroth-Order optimization framework that uses forward passes to fine-tune Large Language Models. |
| Outcome: | The proposed framework achieves 1.7 to 3.0 wall-clock acceleration on LLaMA and OPT models. |
AdaptEvolve: Improving Efficiency of Evolutionary AI Agents through Adaptive Model Selection (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing routing strategies rely on static heuristics or external controllers to optimize performance. |
| Approach: | They propose a framework that leverages intrinsic generation confidence to estimate solvability. |
| Outcome: | Empirical results show that confidence-driven selection yields favorable Pareto frontier . computational cost of state-of-the-art large language models remains a key barrier to scalable deployment . |
Confidence-Weighted Token Set Cover for Early Hypothesis Pruning in Self-Consistency (2026.findings-acl)
Copied to clipboard
| Challenge: | Despite its simplicity and efficacy, the high token expenditure of self-consistency can limit its practical utility. |
| Approach: | They propose a fast weighted set cover algorithm that utilizes two lightweight indicators to prune intermediate hypotheses periodically. |
| Outcome: | The proposed method reduces self-consistency token efficiency by 10-35% on three math benchmarks. |
Data Efficient RLVR via Off-Policy Influence Guidance (2026.acl-long)
Copied to clipboard
Erle Zhu, Dazhi Jiang, Yuan Wang, Xujun Li, Jiale Cheng, Yuxian Gu, Yilin Niu, Aohan Zeng, Jie Tang, Minlie Huang, Hongning Wang
| Challenge: | Existing data selection methods for RLVR are heuristic-based, lacking theoretical guarantees and generalizability. |
| Approach: | They propose an off-policy influence estimation method that approximates data influence using offline trajectories. |
| Outcome: | The proposed method reduces the computational cost of policy rollouts and improves storage and computation efficiency. |
Accurate and Efficient Statistical Testing for Word Semantic Breadth (2026.acl-long)
Copied to clipboard
| Challenge: | Existing studies on semantic breadth and spread of words have focused on semantic relatedness, but contextualized token embeddings can be misleading. |
| Approach: | They propose a Householder-aligned permutation test to isolate dispersion differences from directional differences. |
| Outcome: | The proposed method reduces Type-I error by 32.5% while preserving sensitivity to genuine breadth differences. |