Papers by Jia Li
RealTalk-CN: A Realistic Chinese Speech Task-Oriented Dialogue Benchmark with Cross-Modal Analysis (2026.acl-long)
Copied to clipboard
| Challenge: | Recent advances in speech large language models have enabled end-to-end spoken interactions, but their robustness in real-world applications remains unclear. |
| Approach: | They propose a multi-turn, multi-domain speech–text TOD dataset for Chinese users . it contains 5.4k dialogues with annotations for dialogue states, disfluency types, speaker characteristics . |
| Outcome: | The proposed model can be used to evaluate speech large language models in real-world scenarios . the proposed model is based on 5.4k real human-to-human dialogues with annotations . |
Incentivizing In-depth Reasoning over Long Contexts with Process Advantage Shaping (2026.findings-acl)
Copied to clipboard
| Challenge: | Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective in enhancing LLMs’ short-context reasoning but falters in long-contemporal scenarios requiring precise grounding and multi-hop reasoning. |
| Approach: | They propose a framework that constructs high-difficulty, multi-hop long-context QA pairs with inherent reasoning chains to overcome this bottleneck. |
| Outcome: | The proposed framework outperforms RLVR baselines and matches frontier LLMs while using far fewer parameters. |
Towards Self-Evolving Agents: Enabling Autonomy through Interactive Experience Refinement (2026.findings-acl)
Copied to clipboard
Cheng Yang, Xuemeng Yang, Licheng Wen, Daocheng Fu, Jianbiao Mei, Rong Wu, Pinlong Cai, Yufan Shen, Nianchen Deng, Jia Xu, Botian Shi, Yu Qiao, Haifeng Li
| Challenge: | Large Language Models struggle with complex, multi-step operational tasks because they remain static during inference and cannot learn from past experience. |
| Approach: | They propose a framework that organizes cross-domain insights to facilitate orchestration of long-horizon workflows. |
| Outcome: | The proposed framework outperforms existing methods on the TAC productivity benchmark and shows strong cross-task transferability. |
Enhancing Rumor Detection Methods with Propagation Structure Infused Language Model (2025.coling-main)
Copied to clipboard
| Challenge: | Pretrained Language Models excel in various Natural Language Processing tasks, but performance on social media applications like rumor detection remains suboptimal. |
| Approach: | They propose a pretraining strategy to infuse information from propagation structures into pretrained language models to capture interactions of stance and sentiment crucial for rumor detection. |
| Outcome: | The proposed model outperforms existing methods on social media applications and significantly improves rumor detection performance. |
MTA:A Merge-then-Adapt Framework for Personalized Large Language Models (2026.acl-long)
Copied to clipboard
Xiaopeng Li, Yuanjin Zheng, Wanyu Wang, Wenlin Zhang, Pengyue Jia, Yingyi Zhang, Haiying He, Mengyang Ma, Yiqi Wang, Maolin Wang, Xuetao Wei, Xiangyu Zhao
| Challenge: | Personalized Large Language Models (PLLMs) aim to align outputs with individual user preferences . current methods of fine-tuning a separate module for each user are unscalable . |
| Approach: | They propose a Merge-then-Adapt framework for Personalized Large Language Models . they construct a shared Meta-LoRA bank and propose an Adaptive LoRA Fusion stage . |
| Outcome: | The proposed framework outperforms existing SOTA methods on the LaMP benchmark. |
Data Selection Curriculum for Abstractive Text Summarization (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Abstractive Text Summarization (ATS) models are commonly trained using large-scale data that is randomly shuffled. |
| Approach: | They propose a data selection curriculum scoring system that measures the learning difficulty of an ATS model and expected performance on an instance. |
| Outcome: | The proposed system surpasses baselines on CNN/DailyMail dataset, utilizing 20% of available instances. |
RoadMapper: A Multi-Agent System for Roadmap Generation of Solving Complex Research Problems (2026.findings-acl)
Copied to clipboard
Jiacheng Liu, Zichen Tang, Zhongjun Yang, Xinyi Hu, Xueyuan Lin, Linwei Jia, Ruofei Bai, Rongjin Li, Shiyao Peng, Haocheng Gao, Haihong E
| Challenge: | Existing tools to generate structured content for research tasks are limited in their ability to generate high-quality roadmaps. |
| Approach: | They propose a benchmark to evaluate the ability of large language models (LLMs) to generate high-quality roadmaps for solving complex research problems. |
| Outcome: | The proposed system can improve LLMs’ ability for roadmap generation while saving 84% of the time required by human experts. |
Zero-Shot Cross-Lingual Transfer of Neural Machine Translation with Multilingual Pretrained Encoders (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing work on improving cross-lingual transferability of NMT model is under-explored. |
| Approach: | They propose a model that leverages a multilingual pretrained encoder to improve cross-lingual transferability. |
| Outcome: | The proposed model outperforms mBART and m2m-100 on a zero-shot cross-lingual transfer task. |
Contrastive Preference Learning for Neural Machine Translation (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing discrepancies between token-level objective and overall sequence-level quality of a model are causing exposure bias and other issues in NMT. |
| Approach: | They propose a contrastive preference model that integrates an indicator function to fine-tune a pre-trained model in Neural Machine Translation. |
| Outcome: | The proposed model outperforms the traditional Plackett-Luce model on three language pairs and also outperFORMs token-level and sequence-level baseline models. |
Retrieval and Reasoning on KGs: Integrate Knowledge Graphs into Large Language Models for Complex Question Answering (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have performed impressively in various NLP tasks, but their inherent hallucination phenomena severely challenge their credibility in complex reasoning. |
| Approach: | They propose to integrate explainable Knowledge Graphs (KGs) with LLMs to alleviate hallucinations . they construct subgraphs to enhance the retrieval capabilities of KGs via CoT reasoning. |
| Outcome: | Extensive experiments on two KGQA datasets show that the proposed model achieves convincing performance compared to strong baselines. |
One QuantLLM for ALL: Fine-tuning Quantized LLMs Once for Efficient Deployments (2025.acl-long)
Copied to clipboard
| Challenge: | Quantization has shown promise for Large Language Models, but current methods require lengthy training to alleviate quantization loss. |
| Approach: | They propose to decouple weights and incorporate Low-Rank adapters to reduce weight sharing . they validate the approach on LLaMA2 families and Mistral on downstream evaluation . |
| Outcome: | The proposed approach shows high performance while reducing deployment time faced with multiple scenarios. |
Selective Differential Privacy for Language Modeling (2022.naacl-main)
Copied to clipboard
| Challenge: | Existing methods to protect sensitive data from leaking are over-pessimistic and undifferentiated. |
| Approach: | They propose a new privacy notion, selective differential privacy, to provide rigorous privacy guarantees on the sensitive portion of the data to improve model utility. |
| Outcome: | The proposed privacy-preserving mechanism achieves better utility while remaining safe under various privacy attacks compared to baselines. |
DevEval: A Manually-Annotated Code Generation Benchmark Aligned with Real-World Code Repositories (2024.findings-acl)
Copied to clipboard
Jia Li, Ge Li, Yunfei Zhao, Yongmin Li, Huanyu Liu, Hao Zhu, Lecheng Wang, Kaibo Liu, Zheng Fang, Lanshen Wang, Jiazheng Ding, Xuanming Zhang, Yuqi Zhu, Yihong Dong, Zhi Jin, Binhua Li, Fei Huang, Yongbin Li, Bin Gu, Mengfei Yang
| Challenge: | Existing benchmarks are poorly aligned with real-world code repositories and are insufficient to evaluate the coding abilities of Large Language Models (LLMs). |
| Approach: | They propose a repository-level benchmark named DevEval to evaluate LLMs' coding abilities in real-world code repositories. |
| Outcome: | The proposed benchmarks show that the LLMs perform better in real-world code repositories than existing benchmarks. |
SelfRACG: Enabling LLMs to Self-Express and Retrieve for Code Generation (2025.emnlp-main)
Copied to clipboard
Qian Dong, Jia Chen, Qingyao Ai, Hongning Wang, Haitao Li, null Yiwu, Yao Hu, Yiqun Liu, Shaoping Ma
| Challenge: | Existing retrieval-augmented code generation methods fail to accurately fetch the knowledge required for code generation for consecutive code fragments. |
| Approach: | They propose a paradigm that enables large language models to Self-express their information needs to enhance retrieval-augmented code generation methods. |
| Outcome: | Experiments show that SelfRACG can retrieve external knowledge that better aligns with the LLM’s own information needs, resulting in superior generation performance compared to vanilla RACG. |
Reward Alignment Optimization: A Direct Point-wise Alignment Approach (2026.acl-long)
Copied to clipboard
| Challenge: | Existing Direct Alignment Algorithms (DAAs) are limiting in generalizaiton to implicit rewards. |
| Approach: | They propose a point-wise direct alignment method that uses an explicit reward model to specify exact target generation probabilities and align the policy offline towards them. |
| Outcome: | The proposed method outperforms existing direct alignment algorithms while enabling controllable target probability distributions. |
I2E: From Image Pixels to Actionable Interactive Environments for Text-Guided Image Editing (2026.acl-long)
Copied to clipboard
Jinghan Yu, Junhao Xiao, Chenyu Zhu, Jiaming Li, Jia Li, HanMing Deng, Xirui Wang, Guoli Jia, Jianjun Li, Xiang Bai, Bowen Zhou, Zhiyuan Ma
| Challenge: | Existing text-guided image editing methods rely on end-to-end pixel-level inpainting paradigm . existing models lack such intermediate representations and Reasoning-then-action process . |
| Approach: | They propose a "Decompose-then-Action" paradigm that revisits image editing as an actionable interaction process within a structured environment. |
| Outcome: | The proposed paradigm outperforms existing methods in compositional editing tasks. |
AQE: Argument Quadruplet Extraction via a Quad-Tagging Augmented Generative Approach (2023.findings-acl)
Copied to clipboard
| Challenge: | Argument mining involves multiple subtasks, but each one is insufficient for understanding argumentative structure and reasoning process. |
| Approach: | They propose a quadruplet extraction task that extracts four argumentative components . they use a generative quadragging module to augment the training of the generative framework . |
| Outcome: | The proposed method can extract arguments from a large-scale dataset. |
GeoArena: Evaluating Open-World Geographic Reasoning in Large Vision-Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Existing evaluation paradigms for geographic reasoning are outcome-centric and focus on label matching, leaving the underlying linguistic reasoning chains as unexamined black boxes. |
| Approach: | They propose a dynamic, human-preference-based evaluation framework for benchmarking open-world geographic reasoning. |
| Outcome: | The proposed framework reframes evaluation as a pairwise reasoning alignment task on in-the-wild images, where human judges compare model-generated explanations based on reasoning quality, evidence synthesis, and plausibility. |
MMA: Cross-Domain Knowledge Integration via Mixture of Multi-Domain Agents (2025.findings-emnlp)
Copied to clipboard
| Challenge: | achieving synergistic improvements between generalization and domain specialization remains a challenge in pre-training and post-training. |
| Approach: | They propose a test-time cross-domain knowledge integration method that integrates general-purpose and domain-specific models to enhance their performance on complex, domainspecific tasks. |
| Outcome: | The proposed method combines the outputs of general-purpose and domain-specific models to improve their performance on complex, domainspecific tasks. |
Bridging Relevance and Reasoning: Rationale Distillation in Retrieval-Augmented Generation (2025.findings-acl)
Copied to clipboard
Pengyue Jia, Derong Xu, Xiaopeng Li, Zhaocheng Du, Xiangyang Li, Yichao Wang, Yuhao Wang, Qidong Liu, Maolin Wang, Huifeng Guo, Ruiming Tang, Xiangyu Zhao
| Challenge: | Existing approaches to rerank and align documents based on reasoning capabilities of large language models (LLMs) . prior work shows that LLMs have exceptional reasoning and text generation capabilities . |
| Approach: | They propose a rationale extraction method that leverages reasoning capabilities of large language models to extract the rationales necessary for answering a query. |
| Outcome: | The proposed method is compared with baseline methods on two tasks across three datasets. |
ReTRE: Benchmarking LLM Transfer Robustness with Structure-Preserving Variants (2026.acl-long)
Copied to clipboard
| Challenge: | Learning transfer theory emphasizes that applying acquired knowledge to novel manifestations is a key signal of deep understanding |
| Approach: | They propose a benchmark that probes transfer robustness along two rewrite levels: Near Transfer and Far Transfer. |
| Outcome: | The proposed benchmark demonstrates that large language models are robust when faced with novel manifestations of the same problem. |
RealChart2Code: Bridging the Gap in Real-World Chart-to-Code Generation via Multi-Task Evaluation (2026.acl-long)
Copied to clipboard
Jiajun Zhang, Yuying Li, Zhixun Li, Xingyu Guo, Jingzhuo Wu, Leqi Zheng, Yiran Yang, Jianke Zhang, Qingbin Li, Shannan Yan, Changguo Jia, Junfei Wu, Zilei Wang, Qiang Liu, Liang Wang
| Challenge: | Vision-Language Models (VLMs) have demonstrated impressive capabilities in code generation across various domains, but their ability to replicate complex, multi-panel visualizations remains largely unassessed. |
| Approach: | They propose a large-scale benchmark to evaluate chart generation from large- scale raw data and assess iterative code refinement in a multi-turn conversational setting. |
| Outcome: | The new benchmark evaluates 14 leading VLMs on real-world data and shows they struggle with complex plot structures and authentic data. |
Demystifying Synthetic Data in LLM Pre-training: A Systematic Study of Scaling Laws, Benefits, and Pitfalls (2025.emnlp-main)
Copied to clipboard
Feiyang Kang, Newsha Ardalani, Michael Kuchnik, Youssef Emad, Mostafa Elhoushi, Shubhabrata Sengupta, Shang-Wen Li, Ramya Raghavendra, Ruoxi Jia, Carole-Jean Wu
| Challenge: | a large-scale empirical study compares natural web data, diverse synthetic types, and mixtures of natural and synthetic data. |
| Approach: | They conduct a large-scale empirical study on large-volume LLMs using a unified protocol and scaling laws. |
| Outcome: | The proposed method is faster than pre-training on natural web data, the authors show . their results are consistent with previous studies on rephrased text and textbooks . |
LLM Jailbreak Detection for (Almost) Free! (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for detecting jailbreak prompts entail significant computational costs . |
| Approach: | They propose a free jailbreak detection method which scales logits by temperature to detect jailbreak prompts . |
| Outcome: | The proposed method detects jailbreak prompts with no additional computational costs. |
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 . |
GAN Driven Semi-distant Supervision for Relation Extraction (N19-1)
Copied to clipboard
| Challenge: | Existing methods for relation extraction are limited to costly hand-labeled training sets and hard to be extended to large-scale relations. |
| Approach: | They propose a semi-distant supervision approach for relation extraction by constructing a small accurate dataset and properly leveraging numerous instances without relation labels. |
| Outcome: | The proposed approach achieves significant improvements over baselines on real-world datasets. |
Active Testing: An Unbiased Evaluation Method for Distantly Supervised Relation Extraction (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for distantly supervised relation extraction suffer from low quality of test set, which leads to considerable biased performance evaluation. |
| Approach: | They propose a method to evaluate distantly supervised relation extraction using noisy test sets and manual annotations. |
| Outcome: | Experiments on a widely used benchmark show that the proposed method can yield approximately unbiased evaluations for distantly supervised relation extractors. |
E-ConvRec: A Large-Scale Conversational Recommendation Dataset for E-Commerce Customer Service (2022.lrec-1)
Copied to clipboard
Meihuizi Jia, Ruixue Liu, Peiying Wang, Yang Song, Zexi Xi, Haobin Li, Xin Shen, Meng Chen, Jinhui Pang, Xiaodong He
| Challenge: | Recent research has focused on developing conversational recommendation system (CRS), which provides valuable recommendations to users through conversations. |
| Approach: | They construct an authentic Chinese dialogue dataset consisting of over 25k dialogues and 770k utterances, which contains user profile, product knowledge base, and multiple sequential real conversations between users and recommenders. |
| Outcome: | The proposed dataset contains user profile, product knowledge base, and multiple sequential real conversations between users and recommenders. |
DeepSynth-Eval: Objectively Evaluating Information Consolidation in Deep Survey Writing (2026.findings-acl)
Copied to clipboard
Hongzhi Zhang, Yuanze Hu, Tinghai Zhang, Jia Fu, Tao Wang, Junwei Jing, Zhaoxin Fan, Wei Bi, Ruiming Tang, Han Li, Guorui Zhou, Kun Gai
| Challenge: | Large Language Models (LLMs) are evolving towards autonomous agents . retrieval capabilities are well-benchmarked, but post-retrieval synthesis is under-evaluated due to open-ended writing. |
| Approach: | They propose a benchmark to evaluate information consolidation capabilities using survey papers as gold standards. |
| Outcome: | The proposed benchmark analyzes the post-retrieval synthesis stage of large language models . it leverages high-quality survey papers as gold standards and reverse-engineers research requests . the proposed benchmark outperforms single-turn generation and reduces hallucinations . |
Self-Edit: Fault-Aware Code Editor for Code Generation (2023.acl-long)
Copied to clipboard
| Challenge: | Existing Large language models (LLMs) have low pass rates and accuracy on competitive programming tasks. |
| Approach: | They propose a generate-and-edit approach that uses execution results of generated code from LLMs to improve code quality on competitive programming tasks. |
| Outcome: | The proposed method improves pass@1 by 89% on APPS-dev, 31% on apps-test, and 48% on HumanEval over nine popular code generation LLMs with parameter sizes ranging from 110M to 175B. |
Rethinking Depression Prediction from a Fine-Grained Subscore Modeling Perspective via Multi-Task Learning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for depression assessment rely on standardized ratings, but they are time-consuming and subject to inter-rater variability. |
| Approach: | They propose a fine-grained model for subscore prediction via multi-task learning that can be used to predict depression severity using multiple tasks. |
| Outcome: | The proposed model outperforms baselines and Qwen3-14B direct scoring on the public E-DAIC dataset and to a large-scale private clinical dataset. |
Beyond Fully Random Masking: Attention-Guided Denoising and Optimization for Diffusion Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for full-attention dLLMs rely on random masking strategies that overlook intrinsic token dependencies. |
| Approach: | They propose an attention-guided denoising and optimization framework that aligns training and optimization with attention-derived dependencies. |
| Outcome: | The proposed framework outperforms state-of-the-art methods on mathematical and coding benchmarks. |
CodeAgent: Enhancing Code Generation with Tool-Integrated Agent Systems for Real-World Repo-level Coding Challenges (2024.acl-long)
Copied to clipboard
| Challenge: | Large Language Models excel in simple tasks such as generating standalone code units, but real-world software development often involves complex code repositories with complex dependencies and extensive documentation. |
| Approach: | They propose a novel LLM-based agent framework that employs external tools for effective repo-level code generation. |
| Outcome: | The proposed framework outperforms commercial products like Github Copilot in the humanEval benchmark and shows that it is adaptable and efficient across multiple code generation tasks. |
Table-as-Search: Agentic Information Seeking is Table Completion (2026.findings-acl)
Copied to clipboard
Tian Lan, Felix Henry, Bin Zhu, Qianghuai Jia, Junyang Ren, Qihang PU, Haijun Li, Longyue Wang, Zhao Xu, Weihua Luo
| Challenge: | Current Information Seeking (InfoSeeking) agents struggle to maintain focus and coherence during long-horizon exploration, as tracking search states within one plain-text context is inherently fragile. |
| Approach: | They propose a structured planning framework that reformulates the InfoSeeking task as a Table Completion task. |
| Outcome: | The proposed framework outperforms state-of-the-art frameworks across three kinds of benchmarks, including multi-agent framework and commercial systems. |
ChatHLS: Towards Systematic Design Automation and Optimization for High-Level Synthesis (2026.acl-long)
Copied to clipboard
| Challenge: | High-Level Synthesis (HLS) is a hardware design tool that can be used to design hardware from C-like languages, but its widespread adoption is limited by strict coding constraints and design-specific optimizations. |
| Approach: | They propose a multi-agent HLS design framework that leverages specialized LLMs for automated debugging and directive tuning. |
| Outcome: | The proposed framework outperforms Gemini-3-pro in debugging and speedups across various HLS kernels and neural network accelerators. |
MILL: Mutual Verification with Large Language Models for Zero-Shot Query Expansion (2024.naacl-long)
Copied to clipboard
| Challenge: | Existing methods for query expansion lack corpus-specific knowledge and cost. |
| Approach: | They propose a query-query-document generation method that leverages large language models for mutual verification to produce diverse sub-queries and corresponding documents. |
| Outcome: | The proposed method is fully zero-shot and extensive experiments on three public benchmark datasets demonstrate its effectiveness over existing methods. |
Planning-Guided Tutoring with Assessment-Driven Memory for Pedagogical LLM Tutors (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to simulate tutor behaviors or preferences fail to sustain high-quality pedagogical conversations that provide explicit stepwise scaffolding and adapt to learners’ evolving cognitive states. |
| Approach: | They propose a planning-guided tutoring framework with an assessment-driven memory for multi-turn math dialogue tutoring. |
| Outcome: | Experiments on multi-turn math tutoring benchmarks show that ScaffoldLM significantly improves pedagogical tutoring quality over strong baselines. |
Modularized Interaction Network for Named Entity Recognition (2021.acl-long)
Copied to clipboard
| Challenge: | Named Entity Recognition (NER) models focus on word-level information, while segment-based models focus only on word level information. |
| Approach: | They propose a Modularized Interaction Network (MIN) model which utilizes both word-level information and segment-level dependencies. |
| Outcome: | The proposed model outperforms the current state-of-the-art models on three NER benchmark datasets. |
AlgBench: To What Extent Do Large Reasoning Models Understand Algorithms? (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing benchmarks for algorithmic reasoning fail to answer a critical question: do LRMs master algorithmic thinking? Empirical evaluations on leading LRM models reveal substantial performance heterogeneity, while models perform well on non-optimized tasks, accuracy drops sharply to around 49% on globally optimized algorithms. |
| Approach: | They propose an algorithm-centric benchmark that evaluates large reasoning models under an algorithmic paradigm. |
| Outcome: | Empirical evaluations on leading LRMs reveal substantial performance heterogeneity . models perform well on non-optimized tasks, accuracy drops sharply to around 49% . |
SelfBudgeter: Adaptive Token Allocation for Efficient LLM Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent large reasoning models have shown exceptional performance on various tasks, but they consume excessive tokens even for simple queries, leading to resource waste and prolonged user latency. |
| Approach: | They propose a self-adaptive reasoning strategy that automatically allocates budgets according to problem complexity and introduces GRPO for reinforcement learning to reduce output length. |
| Outcome: | The proposed model achieves an average response length compression of 61% on math reasoning tasks while maintaining accuracy. |
ReflectEvo: Improving Meta Introspection of Small LLMs by Learning Self-Reflection (2025.findings-acl)
Copied to clipboard
Jiaqi Li, Xinyi Dong, Yang Liu, Zhizhuo Yang, Quansen Wang, Xiaobo Wang, Song-Chun Zhu, Zixia Jia, Zilong Zheng
| Challenge: | ReflectEvo-460k is a large-scale, comprehensive, self-generated reflection dataset with broadened instructions and diverse multi-domain tasks. |
| Approach: | They propose a pipeline that iteratively generates self-reflection for self-training and a large-scale reflection dataset with broadened instructions and diverse multi-domain tasks. |
| Outcome: | The proposed pipeline improves Llama-3 reasoning ability by up to 71.2% and Mistral by upto 44.4%. |
Parallelism and Generation Order in Masked Diffusion Language Models: Limits Today, Potential Tomorrow (2026.findings-acl)
Copied to clipboard
Yangyang Zhong, Yanmei Gu, Zhengqing Zang, Xiaomeng Li, Yuqi Ding, Xibei Jia, Yuting Shen, Zhenzhong Lan, Liwang Zhu, Weiping Liu, Junlin Zhou, Haisheng Liu, Zhong Xin Yu, Pengxin Luo, Donglian Qi, Yunfeng Yan, Junbo Zhao
| Challenge: | Autoregressive (AR) language models dominate modern natural language processing due to strong likelihood-based training objectives and reliable left-to-right decoding. |
| Approach: | They characterize MDLM behavior along two dimensions: parallelism strength and generation order . authors propose a Generate-then-Edit paradigm that mitigates dependency loss . |
| Outcome: | The proposed model improves on tasks that require "backward information" the Generate-then-Edit paradigm improves parallel decoding efficiency while reducing dependency loss. |
Natural Response Generation for Chinese Reading Comprehension (2023.findings-emnlp)
Copied to clipboard
| Challenge: | MRC models trained on labeled answers are limited in generating human-like responses in real QA scenarios. |
| Approach: | They construct a dataset called Penguin to promote machine reading comprehension . they use 200k training data with fluent, well-informed responses to train models . |
| Outcome: | The proposed dataset is the first benchmark towards natural response generation in Chinese MRC on a relatively large scale. |
Neural Gibbs Sampling for Joint Event Argument Extraction (2020.aacl-main)
Copied to clipboard
| Challenge: | Existing methods for event argument extraction cannot adequately model the correlation between event arguments and their roles. |
| Approach: | They propose a Bayesian model to jointly extract event arguments using Gibbs sampling . they train two neural networks to model prior distribution and conditional distribution over event arguments . |
| Outcome: | The proposed model can achieve comparable results to existing methods on two widely-used datasets. |
Instance-Guided Prompt Learning for Few-Shot Text Matching (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Few-shot text matching is a more practical technique to determine whether two texts are semantically identical. |
| Approach: | They propose a pluggable prompt learning method for few-shot text matching . they use the semantics of instances to regulate the effects of the gate on the prompt tokens . |
| Outcome: | The proposed method outperforms baselines on MRPC and QQP. |
GenTKG: Generative Forecasting on Temporal Knowledge Graph with Large Language Models (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing methods for temporal relational forecasting are limited and require limited training data. |
| Approach: | They propose a retrieval-augmented generation framework that uses temporal logical rule-based retrieval and parameter-efficient instruction tuning to solve temporal knowledge forecasting challenges. |
| Outcome: | The proposed framework outperforms conventional methods in the temporal knowledge graph domain with low computation resources. |
Chain-of-Thought Prompting Obscures Hallucination Cues in Large Language Models: An Empirical Evaluation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Chain-of-Thought (CoT) prompting can mitigate hallucinations by encouraging step-by-step reasoning, but its impact on halluciation detection remains underexplored. |
| Approach: | They conduct an empirical evaluation of CoT prompting in Large Language Models (LLMs) to examine their impact on hallucination detection methods. |
| Outcome: | The proposed method significantly affects the internal states and token probability distributions of the LLM. |
RelEdit: Evaluating Conceptual Knowledge Editing in Language Models via Relational Reasoning (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing knowledge editing methods struggle to reason about related conceptual knowledge effectively, despite a lack of model-level relational reasoning. |
| Approach: | They propose a benchmark to assess concept-level and instance-level relational reasoning abilities of edited models. |
| Outcome: | The proposed model obtains the best scores on the memory-based in-context editing baseline, MICE, suggesting a promising direction for model editing. |
Joint Learning from Labeled and Unlabeled Data for Information Retrieval (C18-1)
Copied to clipboard
| Challenge: | Recent studies have focused on neural information retrieval (IR) models. |
| Approach: | They propose a framework which can benefit from both labeled and more abundant unlabeled data . they propose supervised retrieval over several strong baselines for IR . |
| Outcome: | The proposed framework can benefit from labeled and more abundant unlabeled data for representation learning in the context of IR. |
Multimodal Fine-grained Context Interaction Graph Modeling for Conversational Speech Synthesis (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods overlook the fine-grained semantic and prosodic interaction modeling at the word level. |
| Approach: | They propose a novel approach to generate conversational prosody by understanding multimodal dialogue history (MDH) using fine-grained semantic and prosodic interaction modeling, they construct specialized multimodal fine-grain dialogue interaction graphs that encode interaction between word-level semantics and prosody. |
| Outcome: | The proposed system outperforms baseline models in terms of prosodic expressiveness. |
Tears or Cheers? Benchmarking LLMs via Culturally Elicited Distinct Affective Responses (2026.acl-long)
Copied to clipboard
Chongyuan Dai, Yaling Shen, Zihan Gao, Jia Li, Yishun Jiang, Yaxiong Wang, Liu Liu, Zongyuan Ge, Jinpeng Hu
| Challenge: | Culture is a fundamental determinant of human affective processing and affective perceptions are often limited by declarative knowledge or established societal customs. |
| Approach: | They propose a multimodal benchmark that leverages LLM-generated provisional labels to isolate cross-cultural emotional distinctions. |
| Outcome: | The proposed benchmark captures cross-cultural emotional distinctions and derives reliable ground-truth annotations through human evaluation. |
VQAGuider: Guiding Multimodal Large Language Models to Answer Complex Video Questions (2025.acl-long)
Copied to clipboard
| Challenge: | Multimodal large language models (MLLMs) can grasp the intention of a question and decomposing it to a series of visual recognition sub-tasks to find out the answer with the help of an agent. |
| Approach: | They propose a framework for multimodal large language models to grasp the intention of a question and decompose it into a series of visual recognition sub-tasks to find out the answer. |
| Outcome: | The proposed framework improves the accuracy of complex video-related questions by 29.6% and 17.2% on CVQA and the existing VQA datasets. |
Unveiling Confirmation Bias in Chain-of-Thought Reasoning (2025.findings-acl)
Copied to clipboard
| Challenge: | Chain-of-thought (CoT) prompting has been widely adopted to enhance the reasoning capabilities of large language models (LLMs). |
| Approach: | They propose to examine how internal beliefs affect reasoning generation and reasoning-guided answer prediction in CoT by decomposing CoT into a two-stage process. |
| Outcome: | The proposed model beliefs affect reasoning generation and reasoning-guided answer prediction in CoT, and the results provide strong evidence of confirmation bias in LLMs. |
LSDC: An Efficient and Effective Large-Scale Data Compression Method for Supervised Fine-tuning of Large Language Models (2025.findings-naacl)
Copied to clipboard
Zhaoguang Long, Yuhao Zhou, Shangqing Zhao, Yupei Ren, Li Cai, Chenghao Jia, Zhe Chen, Zhe Fang, Yuxiang Song, Man Lan
| Challenge: | Large Language Models (LLMs) are expanding in scale and size, increasing computational costs . large-scale data compression techniques can reduce the size of training datasets while maintaining data integrity. |
| Approach: | They propose a large-scale data compression method to reduce the size of training data . they use a bifurcated quantization strategy to maximize the diversity of samples . |
| Outcome: | The proposed method significantly reduces the size of training data while maximizing the submodular gain. |
Logits-Based Finetuning (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for developing compact and efficient large language models lack token-level dependencies and linguistic diversity. |
| Approach: | They propose a logits-based fine-tuning framework that integrates supervised learning and knowledge distillation to build enriched training targets using teacher logits and ground truth labels. |
| Outcome: | The proposed method outperforms existing methods on a large-scale logits dataset and a series of science-focused models. |
Adversarial Attacks on Parts of Speech: An Empirical Study in Text-to-Image Generation (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Recent studies show that text-to-image models are vulnerable to adversarial perturbations . |
| Approach: | They investigate the impact of adversarial attacks on different POS tags within text prompts on T2I models. |
| Outcome: | The proposed model is vulnerable to adversarial perturbations with noun perturbations in text prompts. |
S2R: Teaching LLMs to Self-verify and Self-correct via Reinforcement Learning (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to incentivize LLMs’ deep thinking abilities require large-scale data or significant training efforts. |
| Approach: | They introduce an efficient framework that enhances LLM reasoning by teaching models to self-verify and self-correct during inference. |
| Outcome: | The proposed framework outperforms models trained on long-CoT distilled data with 3.1k initialization samples and achieves an accuracy improvement of 51.0% to 81.6%. |
Exploring Response Uncertainty in MLLMs: An Empirical Evaluation under Misleading Scenarios (2025.emnlp-main)
Copied to clipboard
Yunkai Dang, Mengxi Gao, Yibo Yan, Xin Zou, Yanggan Gu, Jungang Li, Jingyu Wang, Peijie Jiang, Aiwei Liu, Jia Liu, Xuming Hu
| Challenge: | Existing studies have focused mainly on visual–textual misalignment, leaving largely unexplored the MLLMs’ ability to preserve an original correct answer when confronted with misleading information. |
| Approach: | They propose a two-stage evaluation pipeline to quantify the response uncertainty phenomenon by eliciting each model’s original response on unperturbed inputs and injecting explicit (false-answer hints) and implicit (contextual contradictions) misleading instructions. |
| Outcome: | The proposed model overturns a correct answer in 65% of cases after receiving a single deceptive cue. |
Sampling Matters! An Empirical Study of Negative Sampling Strategies for Learning of Matching Models in Retrieval-based Dialogue Systems (D19-1)
Copied to clipboard
| Challenge: | Existing studies focus on constructing a matching model with sophisticated neural architectures, but do little to how to effectively learn such architectures from data. |
| Approach: | They propose to sample negative examples to automatically construct a training set for effective model learning in retrieval-based dialogue systems by using four sampling strategies. |
| Outcome: | The proposed learning method improves the performance of matching models on two benchmarks with three matching models. |
MemSearch-o1: Empowering Large Language Models with Reasoning-Aligned Memory Growth in Agentic Search (2026.acl-long)
Copied to clipboard
Sheng Zhang, Junyi Li, Yingyi Zhang, Pengyue Jia, Yichao Wang, Xiaowei Qian, Wenlin Zhang, Maolin Wang, Yong Liu, Xiangyu Zhao
| Challenge: | Existing methods for memory management struggle to capture fine-grained semantic relations between queries and documents. |
| Approach: | They propose a framework for reasoning and agentic search that grows fine-grained memory fragments from seed tokens from queries, then retraces and deep refines the memory via a contribution function. |
| Outcome: | Experiments on eight benchmark datasets show that MemSearch-o1 significantly mitigates memory dilution and more effectively activates reasoning potential of diverse LLMs. |
EmByte: Decomposition and Compression Learning for Small yet Private NLP (2025.findings-emnlp)
Copied to clipboard
| Challenge: | EMBYTE is a byte-level tokenization model that reduces embedding parameters by up to 94% . it is also resilient to privacy threats such as gradient inversion attacks . |
| Approach: | EMBYTE is a byte-level tokenization model that decomposes subwords into fine-grained byte embeddings and then compresses them via neural projection. |
| Outcome: | EMBYTE achieves substantial embedding compression while preserving accuracy and enhancing privacy. |
QuickLLaMA: Query-aware Inference Acceleration for Large Language Models (2025.coling-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) struggle with capturing long-distance dependencies within sequences to deeply understand semantics. |
| Approach: | They propose a system that captures relevant information within a fixed window size and provides precise answers to queries. |
| Outcome: | The proposed system can read Harry Potter within 30s and accurately answer the questions. |
MMUIE: Massive Multi-Domain Universal Information Extraction for Long Documents (2026.findings-eacl)
Copied to clipboard
| Challenge: | Existing document-level information extraction systems operate at the sentence level or within narrow domains due to annotation constraints. |
| Approach: | They propose a large-scale universal dataset for multi-domain, document-level information extraction from long texts. |
| Outcome: | The proposed dataset integrates traditional knowledge bases with large language models to extract fine-grained entities, aliases, and relation triples across 34 domains. |
Ready Jurist One: Benchmarking Language Agents for Legal Intelligence in Dynamic Environments (2026.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks for legal intelligence are limited to static evaluation paradigms or simplified scenarios. |
| Approach: | They introduce J1-ENVS, the first interactive and dynamic legal environment tailored for LLM-based agents. |
| Outcome: | The proposed framework assesses task performance and procedural compliance across legal proficiency levels. |
How does Misinformation Affect Large Language Model Behaviors and Preferences? (2025.acl-long)
Copied to clipboard
| Challenge: | Existing studies have explored the role of Large Language Models in combating misinformation, but there is still a lack of detailed analysis on the specific aspects and extent to which LLMs are influenced by misinformation. |
| Approach: | They propose to use a benchmark to evaluate LLMs' behavior and knowledge preference toward misinformation to identify their models. |
| Outcome: | The proposed approach is based on 10,346,712 pieces of misinformation and examines knowledge conflicts and stylistic variations. |
CuMA: Aligning LLMs with Sparse Cultural Values via Demographic-Aware Mixture of Adapters (2026.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have a global audience, so alignment must extend to cultural resonance. |
| Approach: | They propose a framework that frames alignment as a conditional capacity separation problem. |
| Outcome: | The proposed framework outperforms both dense baselines and semantic-only MoEs on three large language models. |
Combining Supervised Learning and Reinforcement Learning for Multi-Label Classification Tasks with Partial Labels (2024.acl-long)
Copied to clipboard
| Challenge: | Existing supervised learning methods rely on human annotations, but multi-label tasks pose challenges due to the specific domain knowledge and large class sets. |
| Approach: | They propose a framework that can be used to annotate a subset of positive classes from a multi-label dataset. |
| Outcome: | The proposed framework is generalized and effective across multiple tasks. |
Spec-o3: A Tool-Augmented Vision-Language Agent for Rare Celestial Object Candidate Vetting via Automated Spectral Inspection (2026.acl-long)
Copied to clipboard
| Challenge: | Spec-o3 is a tool-augmented vision-language agent that performs astronomer-aligned spectral inspection. |
| Approach: | They propose a tool-augmented vision-language agent that performs astronomer-aligned spectral inspection via interleaved multimodal chain-of-thought reasoning. |
| Outcome: | Spec-o3 outperforms traditional visual inspection methods on rare-object inspection tasks. |
LLM-empowered Dynamic Prompt Routing for Vision-Language Models Tuning under Long-Tailed Distributions (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Pre-trained vision-language models (VLMs) often suffer from bias in class-imbalanced scenes. |
| Approach: | They propose a multi-dimensional dynamic prompt routing framework that integrates a knowledge base for classes spanning multiple visual-semantic dimensions. |
| Outcome: | The proposed framework achieves comparable results with current SOTA methods on long-tailed benchmarks, including CIFAR-LT, ImageNet-LT and Places-LT. |
Benchmarking Long-Context Language Models on Long Code Understanding (2025.acl-long)
Copied to clipboard
Jia Li, Xuyuan Guo, Lei Li, Kechi Zhang, Ge Li, Jia Li, Zhengwei Tao, Fang Liu, Chongyang Tao, Yuqi Zhu, Zhi Jin
| Challenge: | Currently, long-context language models are limited by the lack of a rigorous evaluation framework for long code understanding. |
| Approach: | They propose to use a long code understanding benchmark LongCodeU to evaluate LCLMs' long code comprehension ability for practical applications. |
| Outcome: | The proposed benchmarks show that current LCLMs are limited in their long code understanding ability, particularly when the long code length is greater than 32K, falling far short of their claimed 128K to 1M context windows. |
Uni-Retrieval: A Multi-Style Retrieval Framework for STEM’s Education (2025.acl-long)
Copied to clipboard
| Challenge: | Current retrieval models focus on natural text-image retrieval, which is insufficient for STEM education contexts due to ambiguities in the retrieval process. |
| Approach: | They propose a diverse expression retrieval task tailored to educational scenarios . they extract query style features as prototypes and build a continuously updated Prompt Bank . |
| Outcome: | The proposed model outperforms existing retrieval models in most retrieval tasks. |
READoc: A Unified Benchmark for Realistic Document Structured Extraction (2025.findings-acl)
Copied to clipboard
Zichao Li, Aizier Abulaiti, Yaojie Lu, Xuanang Chen, Jia Zheng, Hongyu Lin, Xianpei Han, Shanshan Jiang, Bin Dong, Le Sun
| Challenge: | Document Structured Extraction (DSE) is a field of document structure analysis that aims to extract structured content from raw documents. |
| Approach: | They propose a benchmark to evaluate document structured extraction systems by converting unstructured PDFs into semantically rich Markdown. |
| Outcome: | The proposed benchmark is based on 3,576 diverse and real-world documents from arXiv, GitHub, and Zenodo. |
HOTVCOM: Generating Buzzworthy Comments for Videos (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing research focuses on generating descriptive comments in English . hot-comments are important for video marketing and branding, authors say . |
| Approach: | They propose a framework to generate hot-comments on a Chinese video dataset . they use a combination of visual, auditory, and textual data to generate them . |
| Outcome: | The proposed framework shows that it generates hot-comments on both the new and existing datasets. |
DeMAC: Enhancing Multi-Agent Coordination with Dynamic DAG and Manager-Player Feedback (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Multi-agent systems (MAS) powered by large language models struggle to adapt to evolving task dependencies and to handle uncertainties. |
| Approach: | They propose a Dynamic Environment-Aware Manager-Player Agents Coordination framework that enhances multi-agent coordination through long-term strategic planning. |
| Outcome: | The proposed framework outperforms traditional reinforcement learning and human-agent collaboration in the Overcooked simulation. |
CoRec: An Easy Approach for Coordination Recognition (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing syntactic parsers are slow and suffer from errors, especially for long and complicated sentences. |
| Approach: | They propose a pipeline model COordination RECognizer with coordinator identifier and conjunct boundary detector. |
| Outcome: | The proposed model improves the yield of state-of-the-art Open IE models by reducing errors and slow processing time. |
Internal Value Alignment in Large Language Models through Controlled Value Vector Activation (2025.acl-long)
Copied to clipboard
| Challenge: | Existing LLMs do not possess consistent values, but many have been developed to align them at the behavioral level, including supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF). |
| Approach: | They propose a Controlled Value Vector Activation method that directly aligns the internal values of Large Language Models by interpreting how a value is encoded in their latent representations. |
| Outcome: | The proposed method achieves highest success rate across 10 basic values without hurting model performance and fluency, and ensures target values even with opposite and potentially malicious input prompts. |
Can We Trust the Performance Evaluation of Uncertainty Estimation Methods in Text Summarization? (2024.emnlp-main)
Copied to clipboard
| Challenge: | Text summarization is a key natural language generation task, but the high cost of inaccurate summaries raises concerns about the reliability of uncertainty estimation on text summarisation (UE-TS) evaluation methods. |
| Approach: | They propose a UE-TS benchmark that evaluates the uncertainty estimation capabilities of two large language models and one pre-trained language model on three datasets. |
| Outcome: | The proposed benchmark evaluates the uncertainty estimation capabilities of two large language models and one pre-trained language model on three datasets, with human-annotation analysis incorporated where applicable. |
VCSearch: Bridging the Gap Between Well-Defined and Ill-Defined Problems in Mathematical Reasoning (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies have improved the performance of Large language models on well-defined mathematical benchmarks, but they often overlook ill-defined problems. |
| Approach: | They develop a large-scale benchmark that contains over 5,000 ill-defined mathematical problems. |
| Outcome: | The proposed framework improves the accuracy of identifying unsolvable problems by at least 12% across different LLMs, thus achieving stronger robust mathematical reasoning ability. |
Revisiting LoRA through the Lens of Parameter Redundancy: Spectral Encoding Helps (2025.findings-acl)
Copied to clipboard
| Challenge: | Low-Rank Adaptation (LoRA) has emerged as a prominent technique for fine-tuning large foundation models. |
| Approach: | They propose a low-rank Adaptation technique that harnesses the expressiveness of spectral bases to re-parameterize LoRA from a sparse spectral subspace. |
| Outcome: | The proposed technique achieves greater efficiency with fewer parameters than baselines on various downstream tasks, including commonsense reasoning, math reasoning, and code generation. |
Metaphor Detection with Context Enhancement and Curriculum Learning (2024.naacl-long)
Copied to clipboard
| Challenge: | Metaphor detection is a challenging task for natural language processing systems . previous work failed to adequately utilize internal and external semantic relationships . |
| Approach: | They propose a model that leverages the difference between literal and external meanings of words and sentences as the sentence external difference. |
| Outcome: | The proposed model achieves competitive performance across multiple datasets with improved convergence speed compared to other models. |
From Laboratory to Real-World Applications: Benchmarking Agentic Code Reasoning at the Repository Level (2026.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks for repository-level reasoning are inconsistent . repoReason is a white-box diagnostic benchmark centered on abductive assertion verification . |
| Approach: | They propose a white-box diagnostic benchmark centered on abductive assertion verification. |
| Outcome: | The proposed framework eliminates memorization while maintaining authentic logical depth . it also regenerates ground-truth states and quantifyes reasoning via three orthogonal metrics . |
Evaluating the Long-Term Memory of Large Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Recent studies have not thoroughly investigated the memory performance of large language models in long-term tasks. |
| Approach: | They propose a dataset to evaluate the long-term memory capabilities of large language models. |
| Outcome: | The proposed model exhibits memory preferences across different categories of information. |
Focused-DPO: Enhancing Code Generation Through Focused Preference Optimization on Error-Prone Points (2025.findings-acl)
Copied to clipboard
| Challenge: | Current code generation models produce errors concentrated at specific error-prone points, affecting accuracy of code. |
| Approach: | They propose a framework that focuses preference optimization on error-prone areas . focused-DPO improves the accuracy and reliability of code generation by reducing common errors . |
| Outcome: | The proposed framework improves code generation by focusing on error-prone areas. |
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. |
OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis (2025.acl-long)
Copied to clipboard
Qiushi Sun, Kanzhi Cheng, Zichen Ding, Chuanyang Jin, Yian Wang, Fangzhi Xu, Zhenyu Wu, Chengyou Jia, Liheng Chen, Zhoumianze Liu, Ben Kao, Guohao Li, Junxian He, Yu Qiao, Zhiyong Wu
| Challenge: | Graphical User Interface (GUI) agents powered by Vision-Language Models (VLMs) have demonstrated human-like computer control capability. |
| Approach: | They propose a GUI data synthesis pipeline that reverse engineers GUI trajectory construction process by executing pre-defined tasks. |
| Outcome: | The proposed GUI data synthesis pipeline overcomes the bottlenecks of previous methods that rely on pre-defined tasks and limited data diversity. |
EvoCoT: Overcoming the Exploration Bottleneck in Reinforcement Learning for LLMs (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to reinforcement learning with verifiable reward (RLVR) are limited by difficulty or lack of exploration. |
| Approach: | They propose a self-evolving curriculum learning framework based on chain-of-thought reasoning optimization that constrains exploration space by self-generating and verifying CoT trajectories. |
| Outcome: | The proposed framework enables LLMs to solve previously unsolved problems without external supervision and is compatible with various RL fine-tuning methods. |
MultiChartQA: Benchmarking Vision-Language Models on Multi-Chart Problems (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing chart understanding benchmarks focus on single-chart tasks, neglecting multi-hop reasoning required to extract and integrate information from multiple charts. |
| Approach: | They propose a benchmark that evaluates MLLMs’ capabilities in four key areas: direct question answering, parallel question answering and comparative reasoning. |
| Outcome: | The proposed benchmark evaluates MLLMs’ capabilities in four key areas: direct question answering, parallel question answering and comparative reasoning. |
Alleviating Over-smoothing for Unsupervised Sentence Representation (2023.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to learn better unsupervised sentence representations have been successful . over-smoothing problem in unsupervised sentences reduces the capacity of powerful PLMs . |
| Approach: | They propose a method to solve the over-smoothing problem in unsupervised sentence representations by combining negatives from PLMs intermediate layers. |
| Outcome: | The proposed method improves on different strong baselines on Semantic Textual Similarity and Transfer datasets. |
Orca: A Few-shot Benchmark for Chinese Conversational Machine Reading Comprehension (2023.findings-emnlp)
Copied to clipboard
Nuo Chen, Hongguang Li, Junqing He, Yinan Bao, Xinshi Lin, Qi Yang, Jianfeng Liu, Ruyi Gan, Jiaxing Zhang, Baoyuan Wang, Jia Li
| Challenge: | Existing benchmarks for conversational machine reading comprehension are inconsistent with real scenarios. |
| Approach: | They propose to use a Chinese CMRC benchmark to evaluate model's generalization ability towards diverse domains by using zero-shot/few-shot settings. |
| Outcome: | The proposed benchmarks are based on 831 hot-topic driven conversations with 4,742 turns and cover 33 domains. |
Compress to Impress: Unleashing the Potential of Compressive Memory in Real-World Long-Term Conversations (2025.coling-main)
Copied to clipboard
| Challenge: | Existing retrieval-based methods for long-term conversations face challenges in memory database management and accurate memory retrieval, hindering their efficacy in dynamic, real-world interactions. |
| Approach: | They propose a framework that eschews traditional retrieval modules and memory databases and adopts a “One-for-All” approach to manage memory generation, compression, and response generation. |
| Outcome: | The proposed framework produces more nuanced and human-like experiences than retrieval-based methods. |
LegalAgentBench: Evaluating LLM Agents in Legal Domain (2025.acl-long)
Copied to clipboard
Haitao Li, Junjie Chen, Jingli Yang, Qingyao Ai, Wei Jia, Youfeng Liu, Kai Lin, Yueyue Wu, Guozhi Yuan, Yiran Hu, Wuyue Wang, Yiqun Liu, Minlie Huang
| Challenge: | Existing general-domain benchmarks do not capture complexity of real-world judicial cognition and decision-making. |
| Approach: | They propose a benchmark specifically designed to evaluate LLM Agents in the legal domain. |
| Outcome: | The proposed benchmark includes 17 corpora from real-world legal scenarios and provides 37 tools for interacting with external knowledge. |
VisualWebInstruct: Scaling up Multimodal Instruction Data through Web Search (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing vision-language models struggle with reasoning-focused tasks due to the lack of high-quality training data. |
| Approach: | They propose a new approach that leverages search engines to create a multimodal multimodal dataset . they use a set of 30,000 seed images to extract HTML data from 700K unique URLs . |
| Outcome: | The proposed model achieves the best known performance on MMMU-Pro (40.7), MathVerse (42.6), and DynaMath (55.7). |
Decoding Knowledge Attribution in Mixture-of-Experts: A Framework of Basic-Refinement Collaboration and Efficiency Analysis (2025.acl-long)
Copied to clipboard
| Challenge: | Existing attribution methods for dense models fail to capture dynamic routing-expert interactions in sparse MoE architectures. |
| Approach: | They propose to analyze sparse MoE architectures against dense models to capture dynamic routing-expert interactions. |
| Outcome: | The proposed algorithm shows that sparse models achieve higher efficiency per layer . it also shows that deep Qwen-MoE mitigates expert failures while minimizing complexity . |
VoxMind: An End-to-End Agentic Spoken Dialogue System (2026.acl-long)
Copied to clipboard
Tianle Liang, Yifu Chen, Shengpeng Ji, Yijun Chen, Zhiyang Jia, Jingyu Lu, Fan Zhuo, Xueyi Pu, Yangzhuo Li, Zhou Zhao
| Challenge: | Existing research on end-to-end spoken dialogue models has focused on core perception and generation, with limited exploration of tool-augmented extensions. |
| Approach: | They propose a framework to equip end-to-end spoken dialogue models with comprehensive agentic abilities by leveraging a 470-hour AgentChat dataset. |
| Outcome: | The proposed framework outperforms Gemini-2.5-Pro on spoken agent tasks while maintaining general conversational quality. |
A Fused Gromov-Wasserstein Framework for Unsupervised Knowledge Graph Entity Alignment (2023.findings-acl)
Copied to clipboard
| Challenge: | Entity alignment is the task of identifying corresponding entities across different knowledge graphs (KGs). |
| Approach: | They propose an unsupervised entity alignment framework that leverages the Fused Gromov-Wasserstein distance to compare KG semantics and KG structural information. |
| Outcome: | The proposed framework surpasses 21 competitive baselines, including cutting-edge methods, without supervision or hyper-parameter tuning. |
Multi-level Relevance Document Identifier Learning for Generative Retrieval (2025.acl-long)
Copied to clipboard
Fuwei Zhang, Xiaoyu Liu, Xinyu Jia, Yingfei Zhang, Shuai Zhang, Xiang Li, Fuzhen Zhuang, Wei Lin, Zhao Zhang
| Challenge: | Existing methods generate DocIDs based on textual content, which may result in weak semantic connections for similar documents due to variations in expression. |
| Approach: | They propose a new retrieval paradigm that generates unique document identifiers . they propose to use queries as a bridge to connect documents with varying relevance levels . |
| Outcome: | The proposed approach outperforms existing methods on multilingual e-commerce search datasets. |
Semi-automatic Data Enhancement for Document-Level Relation Extraction with Distant Supervision from Large Language Models (2023.emnlp-main)
Copied to clipboard
| Challenge: | Document-level Relation Extraction (DocRE) is a task that aims to extract relations from a long context. |
| Approach: | They propose an automated annotation method that integrates an LLM and a natural language inference module to generate relation triples. |
| Outcome: | The proposed method can extract relations from document-level relation datasets with minimal human effort. |
ConTinTin: Continual Learning from Task Instructions (2022.acl-long)
Copied to clipboard
| Challenge: | a new learning paradigm is proposed for NLP, which seeks supervision for solving a target task. |
| Approach: | They propose a new learning paradigm that uses textual instructions to learn new tasks . the main goal of machine learning algorithms lies in seeking supervision for solving a target task. |
| Outcome: | The proposed learning paradigm is based on a stream of more than 60 tasks . it makes full use of task instructions to improve forward-transfer and backward-transference . |
One Shot Dominance: Knowledge Poisoning Attack on Retrieval-Augmented Generation Systems (2025.findings-emnlp)
Copied to clipboard
Zhiyuan Chang, Mingyang Li, Xiaojun Jia, Junjie Wang, Yuekai Huang, Ziyou Jiang, Yang Liu, Qing Wang
| Challenge: | Existing knowledge poisoning attacks against RAG systems require multiple poisoned documents or can only function effectively on simplistic queries. |
| Approach: | They propose a more realistic knowledge poisoning attack that poisons only a single document while remaining effective for complex multi-hop questions involving complex relationships between multiple elements. |
| Outcome: | The proposed attack achieves success by poisoning only a single document while remaining effective for complex multi-hop questions involving complex relationships between multiple elements. |
Structural Contrastive Pretraining for Cross-Lingual Comprehension (2023.findings-acl)
Copied to clipboard
| Challenge: | Existing methods to train multilingual language models using pretraining tasks like mask language modeling have yielded promising results on a wide range of downstream tasks. |
| Approach: | They propose a new task to align the structural words in a parallel sentence, enhancing models’ ability to comprehend cross-lingual representations. |
| Outcome: | The proposed task improves model's ability to comprehend cross-lingual representations by increasing the frequency of negative pairings. |
In-Context Compositional Generalization for Large Vision-Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent work shows that in-context learning for large language models exhibits compositional generalization capacity. |
| Approach: | They propose a method to exhibit in-context compositional generalization in large vision-language models by combining visual and linguistic modalities. |
| Outcome: | The proposed method reduces redundancy and complexity in in-context learning with LVLMs. |
LegalReasoner: Step-wised Verification-Correction for Legal Judgment Reasoning (2025.acl-long)
Copied to clipboard
Weijie Shi, Han Zhu, Jiaming Ji, Mengze Li, Jipeng Zhang, Ruiyuan Zhang, Jia Zhu, Jiajie Xu, Sirui Han, Yike Guo
| Challenge: | Existing legal judgment prediction methods struggle with logical errors when conducting complex legal reasoning. |
| Approach: | They propose a method which enhances LJP reliability through step-wise verification and correction of the reasoning process. |
| Outcome: | The proposed model significantly improves concordance with court decisions from 72.37 to 80.27 on LLAMA-3.1-70B. |
Evidence Retrieval is almost All You Need for Fact Verification (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing evidence retrieval methods adopt a trivial retrieval strategy, resulting in task-irrelevant evidence and undesirable performance. |
| Approach: | They propose a framework for evidence retrieval and joint fact verification that integrates two modules. |
| Outcome: | The proposed framework improves evidence retrieval and claims verification on a FEVER dataset. |
HSCodeComp: A Realistic and Expert-level Agent Benchmark for Hierarchical Rule Application (2026.acl-long)
Copied to clipboard
| Challenge: | Existing agent benchmarks neglect hierarchical rule application in real-world domains . a critical gap persists in numerous real-life professional domains where decision-making is governed by expert-written rules. |
| Approach: | They propose a benchmark requiring agents to assign a unique 10-digit Harmonized System (HS) Code to products by aligning their fuzzy attributes with strict tariff classification rules. |
| Outcome: | The proposed benchmarks lack hierarchical rule application capability in real-world domains . the proposed benchmark is based on e-commerce and is open-source . |
ControlMath: Controllable Data Generation Promotes Math Generalist Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Currently, mathematical reasoning is one of the most challenging areas for closed-source LLMs. |
| Approach: | They propose an iterative method involving an equation-generator module and two LLM-based agents that generate diverse equations and transform them into math word problems. |
| Outcome: | The proposed method enables the generation of diverse math problems, not limited to specific domains or distributions. |
AirDialogue: An Environment for Goal-Oriented Dialogue Research (D18-1)
Copied to clipboard
| Challenge: | Recent advances in dialogue generation have inspired a number of studies on dialogue systems . however, current datasets are limited in size and the environment for training agents is relatively unsophisticated. |
| Approach: | They propose to use a context-generator to generate travel and flight restrictions to train agents. |
| Outcome: | The proposed model achieves a score of 0.17 while humans can reach 0.91 . the proposed model is based on a large dataset that contains 301,427 goal-oriented conversations . |
WISCA: A Lightweight Model Transition Method to Improve LLM Training via Weight Scaling (2026.findings-acl)
Copied to clipboard
Jiacheng Li, Jianchao Tan, Zhidong Yang, Pingwei Sun, Feiye Huo, Jiayu Qin, Xiangyu Zhang, Maoxin He, Guangming Tan, Weile Jia, Xunliang Cai, Tong Zhao
| Challenge: | Recent advances in training optimization for Transformer-based large language models lack systematic optimization of weight patterns during training. |
| Approach: | They propose a Weight Scaling method that rescales weights while preserving model outputs to improve model training efficiency and model quality. |
| Outcome: | The proposed method significantly improves convergence quality and loss reduction in LLMs with Grouped Query Attention architectures and LoRA fine-tuning tasks. |
Breaking Language Barriers in Multilingual Mathematical Reasoning: Insights and Observations (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing research focuses on developing powerful large language models for mathematical reasoning within monolingual languages. |
| Approach: | They propose to use translation to build powerful multilingual math reasoning models . they propose different training strategies to build xMR LLMs that outperform open-source LLM . |
| Outcome: | The proposed model outperforms open-source LLMs and surpasses ChatGPT in few-shot scenarios. |
MultiDx: A Multi-Source Knowledge Integration Framework towards Diagnostic Reasoning (2026.findings-acl)
Copied to clipboard
Yimin Deng, Zhenxi Lin, Yejing Wang, Guoshuai Zhao, Pengyue Jia, Zichuan Fu, Derong Xu, Yefeng Zheng, Xiangyu Zhao, Li Zhu, Xian Wu, Xueming Qian
| Challenge: | Existing approaches focus on diagnostic reasoning based on internal model knowledge or static knowledge bases. |
| Approach: | They propose a two-stage diagnostic reasoning framework that integrates multi-perspective evidence to generate a diagnostic prediction. |
| Outcome: | The proposed method generates suspected diagnoses and reasoning traces from web search, SOAP-formatted case, and clinical case database. |
H-MAS: Hierarchical Multi-Agent Scheduling for Multi-Tenant LLM Serving (2026.findings-acl)
Copied to clipboard
Yuhan Liu, Cong Xu, Qi Jia, Yihua Wang, Feiyu Chen, Liang Jin, Lu Liu, Yaqian Zhao, Yuting Ding, Xiang Li
| Challenge: | Multi-tenant Model-as-a-Service (MaaS) workloads exhibit non-stationarity across multiple time scales . existing request schedulers often rely on a fixed policy that remains unchanged at runtime . |
| Approach: | They propose a hierarchical multi-agent scheduler that operates in a layered closed loop . they propose to maintain 1.2–3.0 higher Goodput than SGLang and vLLM . |
| Outcome: | Experiments show that H-MAS achieves 1.2–3.0 higher Goodput than SGLang and vLLM . it maintains more stable QoS under diverse request lengths and heterogeneous SLO targets . |
Explicit and Implicit Data Augmentation for Social Event Detection (2025.acl-long)
Copied to clipboard
| Challenge: | Social event detection relies on labeled data, but annotation is costly and labor-intensive. |
| Approach: | They propose a plug-and-play dual augmentation framework that combines explicit text-based and implicit feature-space augmentation to enhance data diversity and model robustness. |
| Outcome: | The proposed framework outperforms the best baseline model by 17.67% on the Twitter2012 dataset and 15.57% on the twitter2018 dataset in terms of the average F1 score. |
Know Thy Enemy: Securing LLMs Against Prompt Injection via Diverse Data Synthesis and Instruction-Level Chain-of-Thought Learning (2026.findings-acl)
Copied to clipboard
Zhiyuan Chang, Mingyang Li, Yuekai Huang, Ziyou Jiang, Xiaojun Jia, Qian Xiong, Junjie Wang, Zhaoyang Li, Qing Wang
| Challenge: | Large language model (LLM)-integrated applications face security vulnerabilities from prompt injection (PI) attacks. |
| Approach: | They propose a model enhancement method that synthesizes diverse training data and employs instruction-level chain-of-thought fine-tuning to enable LLMs to effectively identify and reject malicious instructions regardless of their source or position in the context. |
| Outcome: | The proposed method outperforms baselines in three critical dimensions while maintaining utility performance without degradation. |
KoCo-Bench: Can Large Language Models Leverage Domain Knowledge in Software Development? (2026.acl-long)
Copied to clipboard
Xue Jiang, Ge Li, Jiaru Qian, Xianjie Shi, Chenjie Li, Hao Zhu, Ziyu Wang, Jielun Zhang, Zeyu Zhao, Kechi Zhang, Jia Li, Wenpin Jiao, Zhi Jin, Yihong Dong
| Challenge: | Existing domain-specific code benchmarks focus on assessing what knowledge LLMs possess rather than how they acquire and apply new knowledge. |
| Approach: | They propose a benchmark to evaluate domain specialization methods in real-world software development. |
| Outcome: | KOCO-bench is a new benchmark for evaluating domain specialization methods in real-world software development. |
Beyond the Individual: Virtualizing Multi-Disciplinary Reasoning for Clinical Intake via Collaborative Agents (2026.findings-acl)
Copied to clipboard
Huangwei Chen, Wu Li, Junhao Jia, Yining Chen, Xiaotao Pang, Ya-Long Chen, Li Gonghui, Haishuai Wang, Jiajun Bu, Lei Wu
| Challenge: | Initial outpatient consultations are costly and difficult to scale to real-time intake. |
| Approach: | They propose a synchronous virtual MDT framework that formalizes the consultation state using a structured SOAP representation, separating evidence collection from diagnostic reasoning to improve traceability and bias control. |
| Outcome: | The proposed framework outperforms state-of-the-art models on ClinicalBench and a real-world RAPID-IPN dataset in documentation quality and consultation capability. |
AudioChatLlama: Towards General-Purpose Speech Abilities for LLMs (2024.naacl-long)
Copied to clipboard
Yassir Fathullah, Chunyang Wu, Egor Lakomkin, Ke Li, Junteng Jia, Yuan Shangguan, Jay Mahadeokar, Ozlem Kalinli, Christian Fuegen, Mike Seltzer
| Challenge: | a new model for speech processing and reasoning uses curated data instead of text. |
| Approach: | They extend the instruction-tuned Llama-2 model with end-to-end speech processing and reasoning abilities without using any carefully curated paired data. |
| Outcome: | The proposed model outperforms or outperfects existing models on synthesized and recorded speech QA tests. |
BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Safety backdoors in large language models can be triggered while evading detection during normal interactions. |
| Approach: | They propose a bi-level optimization method that uses a key insight: backdoor triggers induce a uniform drift in the model’s embedding space . inner level identifies universal perturbations to the decoder’s embedded spaces that steer the model towards defender-defined unwanted behaviors; outer level fine-tunes the model to reinforce safe behaviors against these perturbations. |
| Outcome: | The proposed mitigation method reduces the success rate of safety backdoor attacks from over 95% to 1% for general harmful behaviors and from 47% to 0% for Sleeper Agents, without compromising the model’s usefulness. |
TEBNER: Domain Specific Named Entity Recognition with Type Expanded Boundary-aware Network (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to label data and identify entities require large amounts of manually annotated texts for training supervised models. |
| Approach: | They propose a dictionary extension method which extracts new entities through the type expanded model. |
| Outcome: | The proposed method outperforms state-of-the-art supervised systems on different types of datasets and surpasses supervised models. |
VulAgent: Hypothesis-Validation Driven Multi-Agent Architecture for Vulnerability Detection (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent reports indicate that software vulnerabilities caused by insecure coding practices remain a major security threat. |
| Approach: | They propose a multi-agent vulnerability detection framework based on hypothesis validation . they use multi-view analyzers to localize and localize security-sensitive operations . |
| Outcome: | The proposed framework reduces false positives and increases accuracy by 6.6 percentage points on PrimeVul and SVEN. |
ReContraster: Making Your Posters Stand Out with Regional Contrast (2026.acl-long)
Copied to clipboard
| Challenge: | Effective poster design requires rapidly capturing attention and clearly conveying messages. |
| Approach: | They propose a poster-based model that leverages regional contrast to make posters stand out. |
| Outcome: | The proposed model outperforms state-of-the-art methods in producing striking posters. |
Denoising based Sequence-to-Sequence Pre-training for Text Generation (D19-1)
Copied to clipboard
| Challenge: | PoDA pre-trains encoders and decoders by denoising noise-corrupted text . Unlike encoder-only or decode-only methods, it can be used for text generation tasks without using any task-specific techniques. |
| Approach: | They propose a sequence-to-sequence (seq2sequ) pre-training method PoDA which denoises autoencoders by denoising noise-corrupted text. |
| Outcome: | The proposed method improves model performance over strong baselines without using any task-specific techniques and significantly speed up convergence. |
Quantized Side Tuning: Fast and Memory-Efficient Tuning of Quantized Large Language Models (2024.acl-long)
Copied to clipboard
Zhengxin Zhang, Dan Zhao, Xupeng Miao, Gabriele Oliaro, Zhihao Zhang, Qing Li, Yong Jiang, Zhihao Jia
| Challenge: | Existing methods to finetun large language models (LLMs) only update a small number of trainable parameters, or attempt to reduce the memory footprint during the training phase of the finetune process. |
| Approach: | They propose quantized side tuing (QST) which quantizes an LLM’s model weights into 4-bit to reduce the memory footprint of the original weights. |
| Outcome: | The proposed method reduces the memory footprint of the model weights, optimizer states, and intermediate activations while reducing the memory requirements. |
Making RALM Robust to Irrelevant Contexts via Layer Knowledge Guided Attention (2025.findings-acl)
Copied to clipboard
Weijie Shi, Hao Chen, Jiaming Li, Yao Zhao, Yazhong Zhang, Qijin Chen, Jipeng Zhang, Ruiyuan Zhang, Jia Zhu, Jiajie Xu, Xiaofang Zhou
| Challenge: | Large language models (LLMs) face factual hallucination and knowledge obsolescence when tackling knowledge-intensive tasks. |
| Approach: | They propose a layer-knowledge guided attention method which harnesses the layer-wise knowledge of large language models to optimize per-layer attention on useful passages. |
| Outcome: | The proposed method outperforms existing methods on RALM benchmarks. |
Delete, Retrieve, Generate: a Simple Approach to Sentiment and Style Transfer (N18-1)
Copied to clipboard
| Challenge: | Previous work using adversarial methods has struggled to produce high-quality outputs. |
| Approach: | They propose a method that transforms a sentence to alter a specific attribute while preserving its attribute-independent content. |
| Outcome: | The proposed method generates grammatical and appropriate responses on 22% more inputs than the best previous system, averaged over three attribute transfer datasets. |
JailMeter: An Evidence-Based Evaluation Framework for Jailbreak Attacks on Large Language Models (2026.findings-acl)
Copied to clipboard
Qingjia Huang, Jingyu Zhang, Jianguo Wu, Yakai Li, Weijuan Zhang, Yankai Rong, Junyi Yao, Shengzhi Zhang, Xiaoqi Jia
| Challenge: | Currently, evaluation criteria and methods used for jailbreak effectiveness are inconsistent. |
| Approach: | They propose a framework to measure jailbreak effectiveness using a model that filters out jailbreak noise while preserving the original malicious question. |
| Outcome: | The proposed framework outperforms existing evaluation methods on a challenging benchmark containing 330 human-labeled, non-rejected jailbreak instances. |
Golden Touchstone: A Comprehensive Bilingual Benchmark for Evaluating Financial Large Language Models (2025.findings-emnlp)
Copied to clipboard
Xiaojun Wu, Junxi Liu, Huan-Yi Su, Zhouchi Lin, Yiyan Qi, Chengjin Xu, Jiajun Su, Jiajie Zhong, Fuwei Wang, Saizhuo Wang, Fengrui Hua, Jia Li, Jian Guo
| Challenge: | Existing financial benchmarks suffer from limited language and task coverage, low-quality datasets, and inadequate adaptability for LLM evaluation. |
| Approach: | They propose a bilingual benchmark for financial LLMs that assesses models’ language understanding and generation capabilities. |
| Outcome: | The proposed bilingual benchmark assesses models’ language understanding and generation capabilities. |
Reinforced Query Reasoners for Reasoning-intensive Retrieval Tasks (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for retrieval of information excel at textual and semantic matching but struggle in reasoning-intensive retrieval tasks. |
| Approach: | They propose a family of small-scale language models for query reasoning and rewriting in reasoning-intensive retrieval. |
| Outcome: | The proposed model outperforms existing models on a BRIGHT benchmark with BM25 retrievers. |
LIME: Less Is More for MLLM Evaluation (2025.findings-acl)
Copied to clipboard
King Zhu, Qianbo Zang, Shian Jia, Siwei Wu, Feiteng Fang, Yizhi Li, Shuyue Guo, Tianyu Zheng, Jiawei Guo, Bo Li, Haoning Wu, Xingwei Qu, Jian Yang, Ruibo Liu, Xiang Yue, Jiaheng Liu, Chenghua Lin, Hamid Alinejad-Rokny, Min Yang, Shiwen Ni, Wenhao Huang, Ge Zhang
| Challenge: | Existing MLLM benchmarks and unified evaluation frameworks cannot accurately and efficiently reflect the ability of MLMLs. |
| Approach: | They propose a semi-automated benchmark curated using a pipeline that filters out uninformative samples and eliminates answer leakage by focusing on tasks that require image-based understanding. |
| Outcome: | The proposed benchmark reduces the number of samples by 76% and evaluation time by 77% while it can more effectively distinguish different models’ abilities. |
PIPER: Benchmarking and Prompting Event Reasoning Boundary of LLMs via Debiasing-Distillation Enhanced Tuning (2025.acl-long)
Copied to clipboard
Zhicong Lu, Changyuan Tian, PeiguangLi PeiguangLi, Li Jin, Sirui Wang, Wei Jia, Ying Shen, Guangluan Xu
| Challenge: | Existing studies on Large Language Models (LLMs) have failed to evaluate their performance in event reasoning with a single event relational type or reasoning format. |
| Approach: | They propose a benchmark to evaluate LLMs' event reasoning capability using a single event relational type or reasoning format. |
| Outcome: | The proposed model improves on 10K diverse instruction-tuning demonstrations to alleviate event reasoning-oriented data scarcity. |
Logic Jailbreak: Efficiently Unlocking LLM Safety Restrictions Through Formal Logical Expression (2026.findings-acl)
Copied to clipboard
Jingyu Peng, Maolin Wang, Nan Wang, Jiatong Li, Yuchen Li, Yuyang Ye, Wanyu Wang, Pengyue Jia, Kai Zhang, Xiangyu Zhao
| Challenge: | Despite advances in aligning LLMs with human values, current safety mechanisms remain vulnerable to jailbreak attacks. |
| Approach: | They propose a black-box jailbreak method that uses logical expression translation to bypass LLM safety mechanisms. |
| Outcome: | The proposed method exploits the distributional gap between alignment data and logic-expressed inputs while preserving the underlying semantic intent and readability while evading safety constraints. |
Seek-and-Solve: Benchmarking MLLMs for Visual Clue-Driven Reasoning in Daily Scenarios (2026.findings-acl)
Copied to clipboard
Xiaomin Li, Tala Wang, Zichen Zhong, Ying Zhang, Zirui Zheng, Takashi Isobe, Dezhuang Li, Huchuan Lu, You He, Xu Jia
| Challenge: | Existing benchmarks focus on evaluating MLLMs’ pre-existing knowledge or perceptual understanding, often neglecting the critical capability of reasoning. |
| Approach: | They propose a benchmark designed for visual clue-driven reasoning in daily scenarios that combines rigorous grounding in authentic daily activities and challenging query design that necessitates more than surface-level perception. |
| Outcome: | The proposed benchmark identifies visual clues and their ability to provide robust reasoning in daily scenarios. |
Large Language Models Meet Harry Potter: A Dataset for Aligning Dialogue Agents with Characters (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing models that can create open-domain dialogue agents lack character representation and annotations. |
| Approach: | They propose a dataset to study character alignment and character representation . it includes all dialogue sessions from the Harry Potter series and includes annotations . |
| Outcome: | The proposed dataset can be used as a universal benchmark for character-driven LLMs. |
DocAgent: An Agentic Framework for Multi-Modal Long-Context Document Understanding (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to document understanding are limited due to limited context length or fail to fully leverage multi-modal information. |
| Approach: | They propose a multi-agent framework for long-context document understanding that imitates human reading practice. |
| Outcome: | The proposed framework surpasses human-level benchmarks on long-context document understanding while maintaining a short context length. |
Multi-Perspective Context Aggregation for Semi-supervised Cloze-style Reading Comprehension (C18-1)
Copied to clipboard
| Challenge: | Recent studies have shown that cloze-style reading comprehension is a popular task for measuring the progress of natural language understanding. |
| Approach: | They propose a multi-perspective framework which can be seen as joint training of heterogeneous experts and aggregate context information from different perspectives. |
| Outcome: | The proposed framework achieves new state-of-the-art over previous strong baselines on a recently released cloze-test dataset. |
Parameter-Efficient Fine-Tuning via Circular Convolution (2025.findings-acl)
Copied to clipboard
| Challenge: | Low-Rank Adaptation (LoRA) has gained popularity for fine-tuning large foundation models, but its intrinsic low-rank characteristic may limit its performance. |
| Approach: | They propose a low-rank adaptive method that uses low-ranked matrices to represent weight changes. |
| Outcome: | The proposed method reduces trainable parameters and mitigates heavy memory consumption associated with full delta matrices by sequentially multiplying mathbf A and mathbb B with the activation. |