Papers by Yu Hao
Interventional Training for Out-Of-Distribution Natural Language Understanding (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for NLU training use only known and single confounders, but in many NLU tasks the confounder can be unknown and multifactorial. |
| Approach: | They propose a method that performs multi-granular intervention with identified multifactorial confounders by using a bottom-up automatic intervention method. |
| Outcome: | The proposed method performs multi-granular intervention with identified multifactorial confounders on three NLU tasks, namely, natural language inference, fact verification and paraphrase identification. |
ERNIE-Doc: A Retrospective Long-Document Modeling Transformer (2021.acl-long)
Copied to clipboard
| Challenge: | Existing models for document-level language pretraining are not suitable for long documents due to their quadratically increasing memory and time consumption. |
| Approach: | They propose a document-level language pretraining model based on Recurrence Transformers. |
| Outcome: | The proposed model outperforms existing models on language understanding tasks. |
AIDA-SEAT: Towards Reliable AI Doctor Assistant via State-Evaluation-Action Tree Enhanced LLMs in Online Hospital (2026.acl-industry)
Copied to clipboard
Lianxin Sun, Xiaoying Ying, Guangya Yu, Weiyan Zhang, Chenhao Guan, Hao He, Mingxi Shang, Jianhua Li, ChunMing Wang, Tong Ruan
| Challenge: | Existing systems rely on large language models or retrieval-augmented generation (RAG) but these methods lack the explicit logical pathways essential for multi-step reasoning. |
| Approach: | They propose an AIDA-SEAT framework to provide reliable clinical decision-making support by transforming and modifying medical documents and doctors' state-evaluation-action trees. |
| Outcome: | The proposed framework achieves 1.01% higher than current state-of-the-art (SOTA) baselines across five departments, including common RAG-based methods. |
Safety Sidecar: Reflection-Driven Runtime Control for Safer Agents (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing safety controls fail to provide runtime intervention or cross-architecture portability for autonomous LLM agents. |
| Approach: | They propose a model-agnostic, plug-and-play module to provide arbitrary agent safety control and auditability. |
| Outcome: | The proposed module improves the secure-solution rate by 2.9–11.2 percentage points . it adds only 3.2s to end-to-end latency and a negligible average cost of 5.37 10-4 per scenario . |
XLTime: A Cross-Lingual Knowledge Transfer Framework for Temporal Expression Extraction (2022.findings-naacl)
Copied to clipboard
| Challenge: | Temporal Expression Extraction (TEE) is essential for understanding time in natural language. |
| Approach: | They propose a framework for multilingual Temporal Expression Extraction that leverages pre-trained language models to prompt cross-language knowledge transfer from English to non-English languages. |
| Outcome: | The proposed framework outperforms the existing SOTA methods on French, Spanish, Portuguese, and Basque by large margins. |
Pre-training Multi-task Contrastive Learning Models for Scientific Literature Understanding (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Pre-trained language models (LMs) have shown effectiveness in literature understanding tasks, especially when tuned via contrastive learning. |
| Approach: | They propose a multi-task contrastive learning framework that enables common knowledge sharing across different scientific literature understanding tasks while preventing task-specific skills from interfering with each other. |
| Outcome: | The proposed framework outperforms state-of-the-art pre-trained language models on a comprehensive dataset. |
FLAIR: Steering LLM Mathematical Problem Solving based on A Fuzzy-Logic-AssIsted Reasoner (2026.acl-long)
Copied to clipboard
Hao Wu, Hongru Sun, Wanqing Li, Xinguo Yu, Hao Ming, Xiao Luo, Wenbin Zhang, Jiahong Zhao, Yi Guo, Jie Yang
| Challenge: | Existing approaches to mathematical reasoning rely on static heuristics or pre-determined reasoning strategies. |
| Approach: | They propose an adaptive framework that integrates fuzzy theory into LLM-based mathematical reasoning. |
| Outcome: | The proposed framework outperforms state-of-the-art models while offering effective and interpretable diagnostics of intermediate problem-solving states. |
Argumentation Mining on Essays at Multi Scales (2020.coling-main)
Copied to clipboard
| Challenge: | Argumentation mining on essays is a new task in natural language processing. |
| Approach: | They propose a multi-scale argumentation mining model which aims to identify the types and locations of argumentation components from essay text. |
| Outcome: | The proposed model outperforms existing models on mining all types of argumentation components on the Persuasive Essay dataset. |
OpenWebAgent: An Open Toolkit to Enable Web Agents on Large Language Models (2024.acl-demos)
Copied to clipboard
Iat Long Iong, Xiao Liu, Yuxuan Chen, Hanyu Lai, Shuntian Yao, Pengbo Shen, Hao Yu, Yuxiao Dong, Jie Tang
| Challenge: | OpenWebAgent integrates large language models and large multimodal models to improve web automation. |
| Approach: | They propose to integrate large language models and large multimodal models into an open toolkit to optimize web automation. |
| Outcome: | The open toolkit integrates both large language models (LLMs) and large multimodal models (LMMs) it enables the development of powerful, task-oriented web agents, significantly enhancing user experience and operational efficiency on the web. |
Are Large Language Models Reliable Reviewers? A Benchmark for Error Detection in Financial Documents (2026.findings-acl)
Copied to clipboard
Ying He, Zhouhong Gu, Zhecheng Hu, Yubo Zhou, Hao Shen, Jiaqing Liang, Zhaoqian Dai, Ma Shuguang, Fei Yu, Yanghua Xiao, Zhixu Li
| Challenge: | Existing LLMs struggle to identify errors in financial documents, a study shows . 18% of financial practitioners make errors daily, one-third make errors several times weekly, and 59% make errors multiple times monthly. |
| Approach: | They introduce FinED-Bench, a publicly available Benchmark for financial error detection . it covers nine real-world financial scenarios and includes over 900 documents in 2025 . supervised fine-tuning can significantly improve the performance of weaker LLMs, they show . |
| Outcome: | The proposed benchmark covers nine real-world financial scenarios and includes over 900 documents reported in 2025 that are unseen by existing language models. |
Guided Knowledge Generation with Language Models for Commonsense Reasoning (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have achieved notable success in commonsense reasoning tasks, benefiting from extensive world knowledge acquired through extensive pretraining. |
| Approach: | They propose a method to generate knowledge explanations and to automatically assign labels based on the probability of correct answers. |
| Outcome: | The proposed method outperforms baselines on four widely-used commonsense reasoning benchmarks and shows that it can generate high quality knowledge leading to correct answers. |
Task-Oriented Dialogue as Dataflow Synthesis (2020.tacl-1)
Copied to clipboard
Jacob Andreas, John Bufe, David Burkett, Charles Chen, Josh Clausman, Jean Crawford, Kate Crim, Jordan DeLoach, Leah Dorner, Jason Eisner, Hao Fang, Alan Guo, David Hall, Kristin Hayes, Kellie Hill, Diana Ho, Wendy Iwaszuk, Smriti Jha, Dan Klein, Jayant Krishnamurthy, Theo Lanman, Percy Liang, Christopher H. Lin, Ilya Lintsbakh, Andy McGovern, Aleksandr Nisnevich, Adam Pauls, Dmitrij Petters, Brent Read, Dan Roth, Subhro Roy, Jesse Rusak, Beth Short, Div Slomin, Ben Snyder, Stephon Striplin, Yu Su, Zachary Tellman, Sam Thomson, Andrei Vorobev, Izabela Witoszko, Jason Wolfe, Abby Wray, Yuchen Zhang, Alexander Zotov
| Challenge: | Existing approaches to task-oriented dialogue represent dialogue state as a dataflow graph . microsoft's SMCalFlow dataset features complex dialogues about events, weather, places, and people . |
| Approach: | They propose a dataflow graph-based dialogue agent that maps each user utterance to a program that extends this graph. |
| Outcome: | The proposed framework improves representability and predictability in natural dialogues . it uses dataflow graphs and metacomputation to map user intents to a program . |
FOREVER: Forgetting Curve-Inspired Memory Replay for Language Model Continual Learning (2026.acl-long)
Copied to clipboard
Yujie Feng, Hao Wang, Jian Li, Xu Chu, Zhaolu Kang, Yiran Liu, Yasha Wang, Philip S. Yu, Xiao-Ming Wu
| Challenge: | Continual learning (CL) for large language models (LLMs) aims to enable sequential knowledge acquisition without catastrophic forgetting. |
| Approach: | They propose a framework that aligns replay schedules with a model-centric notion of time. |
| Outcome: | Experiments on three benchmarks show that FOREVER consistently mitigates catastrophic forgetting. |
Detecting Subtle Differences between Human and Model Languages Using Spectrum of Relative Likelihood (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for detecting modelgenerated texts from human texts are limited by the fact that absolute likelihood values of texts are bound to certain linguistic and cognitive constraints. |
| Approach: | They propose to use relative likelihood values instead of absolute ones to extract useful features from the spectrum-view of likelihood for the human-model text detection task. |
| Outcome: | The proposed method can reveal subtle differences between human and model languages, which find theoretical roots in psycholinguistics studies. |
LM-Infinite: Zero-Shot Extreme Length Generalization for Large Language Models (2024.naacl-long)
Copied to clipboard
| Challenge: | Currently, large language models (LLMs) train on short text segments due to the computational overhead quadratic in the input lengths of their Transformer architectures. |
| Approach: | They propose a method that allows LLMs pre-trained with 2K or 4K-long segments to generalize to up to 200M length inputs while retaining perplexity. |
| Outcome: | The proposed method achieves 2.7 decoding speed up and 7.5 memory saving over the original model. |
T2I-FactualBench: Benchmarking the Factuality of Text-to-Image Models with Knowledge-Intensive Concepts (2025.acl-long)
Copied to clipboard
Ziwei Huang, Wanggui He, Quanyu Long, Yandi Wang, Haoyuan Li, Zhelun Yu, Fangxun Shu, Weilong Dai, Hao Jiang, Fei Wu, Leilei Gan
| Challenge: | Existing studies on text-to-image (T2I) models focus on text alignment, image quality, and object composition capabilities. |
| Approach: | They propose a T2I-FactualBench benchmark to evaluate the factuality of knowledge-intensive concept generation. |
| Outcome: | The proposed framework evaluates the factuality of knowledge-intensive concept generation tasks. |
Exploiting the Index Gradients for Optimization-Based Jailbreaking on Large Language Models (2025.coling-main)
Copied to clipboard
| Challenge: | Despite advances in training Large Language Models, they remain vulnerable to jailbreak, an adversarial attack method. |
| Approach: | They propose an adversarial jailbreak algorithm that exploits the gradient information of the suffix tokens to accelerate the optimization process. |
| Outcome: | The proposed model achieves 1.5x speedup while maintaining high attack success rates. |
Align2LLaVA: Cascaded Human and Large Language Model Preference Alignment for Multi-modal Instruction Curation (2025.findings-acl)
Copied to clipboard
Hongzhe Huang, Jiang Liu, Zhewen Yu, Li Cai, Dian Jiao, Wenqiao Zhang, Siliang Tang, Juncheng Li, Hao Jiang, Haoyuan Li, Yueting Zhuang
| Challenge: | Recent advances in Multi-modal Large Language Models (MLLMs) introduce significant variability in data quality. |
| Approach: | They propose to use human and LLM preference alignment to compress large corpus of machine-generated multimodal instructions into a compact and high-quality form. |
| Outcome: | The proposed algorithm outperforms LLaVA-series models in MLLM benchmarks by 90% . it uses human and LLM preference alignment to compress a large dataset . |
Rethinking Text-to-SQL: Dynamic Multi-turn SQL Interaction for Real-world Database Exploration (2026.findings-acl)
Copied to clipboard
Linzhuang Sun, Tianyu Guo, Hao Liang, Ruitong Liu, Yuying Li, Qifeng Cai, Jingxuan Wei, Yuchen Wu, Bihui Yu, Xiangxiang Zhang, Wentao Zhang, Bin Cui
| Challenge: | Structured Query Language (SQL) is the cornerstone for data-driven decision-making. |
| Approach: | They propose a benchmark to rigorously evaluate Large Language Models within a dynamic interaction framework. |
| Outcome: | The proposed benchmark aims to rigorously evaluate LLMs within a dynamic interaction framework. |
TeamLoRA: Boosting Low-Rank Adaptation with Expert Collaboration and Competition (2025.acl-long)
Copied to clipboard
Tianwei Lin, Jiang Liu, Wenqiao Zhang, Yang Dai, Haoyuan Li, Zhelun Yu, Wanggui He, Juncheng Li, Jiannan Guo, Hao Jiang, Siliang Tang, Yueting Zhuang
| Challenge: | Existing methods for fine-tuning are resource-efficient, but performance often falls short . a new approach, TeamLoRA, integrates collaborative and competitive modules to improve performance. |
| Approach: | They propose to introduce task-specific LoRA as domain experts to improve learning efficiency . teamLoRA integrates collaborative and competition modules to improve model learning . |
| Outcome: | Experiments show that TeamLoRA improves performance in multi-task learning . teamLorea integrates collaborative and competitive modules to improve performance . |
SOTOPIA-π: Interactive Learning of Socially Intelligent Language Agents (2024.acl-long)
Copied to clipboard
Ruiyi Wang, Haofei Yu, Wenxin Zhang, Zhengyang Qi, Maarten Sap, Yonatan Bisk, Graham Neubig, Hao Zhu
| Challenge: | Existing studies on building language agents have not addressed this social learning gap. |
| Approach: | They propose an interactive learning method that improves the social intelligence of language agents by using behavior cloning and self-reinforcement based training on filtered social interaction data. |
| Outcome: | The proposed method allows a 7B LLM to reach the social goal completion ability of an expert model (GPT-4-based agent) without the loss of more generic abilities, such as the ability to answer knowledge-based questions. |
The Microsoft Toolkit of Multi-Task Deep Neural Networks for Natural Language Understanding (2020.acl-demos)
Copied to clipboard
Xiaodong Liu, Yu Wang, Jianshu Ji, Hao Cheng, Xueyun Zhu, Emmanuel Awa, Pengcheng He, Weizhu Chen, Hoifung Poon, Guihong Cao, Jianfeng Gao
| Challenge: | MT-DNN is an open-source natural language understanding toolkit . it allows researchers and developers to train customized deep learning models . |
| Approach: | They present MT-DNN, an open-source natural language understanding toolkit . it is designed to facilitate rapid customization for a broad spectrum of NLU tasks . MT supports multi-task knowledge distillation, which can substantially compress a deep neural model without significant performance drop. |
| Outcome: | The proposed model can significantly compress a large model without significant performance drop. |
ReEval: Automatic Hallucination Evaluation for Retrieval-Augmented Large Language Models via Transferable Adversarial Attacks (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing static benchmarks do not guarantee that models can use the provided evidence for answering, which is essential to avoid hallucination when the required knowledge is new or private. |
| Approach: | They propose to automatically perturb existing static one for dynamic evaluation by using a chatGPT framework and a set of open-domain QA datasets. |
| Outcome: | The proposed framework generates new test cases on two open-domain QA datasets and is human-readable and useful to trigger hallucination in LLMs. |
CLEAR: A Framework Enabling Large Language Models to Discern Confusing Legal Paragraphs (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing work focuses on enabling LLMs to leverage legal rules to tackle complex legal reasoning tasks, but ignores their ability to understand legal rules. |
| Approach: | They propose a legal paragraph prediction task that aims to predict the legal paragraph given criminal facts and a framework CLEAR to enhance their legal reasoning ability. |
| Outcome: | The proposed model improves the ability of LLMs to analyze legal cases with the guidance of legal rule insights. |
CB-Whisper: Contextual Biasing Whisper Using Open-Vocabulary Keyword-Spotting (2024.lrec-main)
Copied to clipboard
Yuang Li, Yinglu Li, Min Zhang, Chang Su, Jiawei Yu, Mengyao Piao, Xiaosong Qiao, Miaomiao Ma, Yanqing Zhao, Hao Yang
| Challenge: | End-to-end automatic speech recognition systems struggle to recognize rare name entities such as personal names, organizations and terminologies that are not frequently encountered in the training data. |
| Approach: | They propose a convolutional neural network-based ASR system that performs open-vocabulary keyword-spotting before the decoder to match the features between the entities and the utterances. |
| Outcome: | The proposed system significantly improves mixed-error-rate (MER) and entity recall compared to the original Whisper model on three internal datasets and two publicly available datasets. |
ReasonMed: A 370K Multi-Agent Generated Dataset for Advancing Medical Reasoning (2025.emnlp-main)
Copied to clipboard
Yu Sun, Xingyu Qian, Weiwen Xu, Hao Zhang, Chenghao Xiao, Long Li, Deli Zhao, Wenbing Huang, Tingyang Xu, Qifeng Bai, Yu Rong
| Challenge: | Existing medical reasoning datasets are limited in scale and typically rely on incomplete data. |
| Approach: | They propose to use ReasonMed to train medical reasoning models using a multi-agent generation, verification, and refinement pipeline. |
| Outcome: | The largest medical reasoning dataset to date surpasses the prior best sub-10B models by 4.17% and even exceeds LLaMA3.1-70B on PubMedQA by 4.60%. |
Generating Sentences from Disentangled Syntactic and Semantic Spaces (P19-1)
Copied to clipboard
| Challenge: | Variational auto-encoders (VAEs) are widely used in natural language generation due to the regularization of the latent space. |
| Approach: | They propose to generate sentences from disentangled syntactic and semantic spaces by using the linearized tree sequence. |
| Outcome: | The proposed method achieves similar or better performance in various tasks compared with state-of-the-art models. |
Retrievals Can Be Detrimental: Unveiling the Backdoor Vulnerability of Retrieval-Augmented Diffusion Models (2026.acl-long)
Copied to clipboard
| Challenge: | Retrieval-augmented diffusion models (RDMs) have been developed to enhance performance with reduced parameters. |
| Approach: | They propose to integrate retrieval-augmented diffusion models with Retrieval-augmented generation (RAG) that enhances performance with reduced parameters. |
| Outcome: | The proposed framework achieves outstanding attack effects while maintaining benign utility. |
Compositional Generalization for Neural Semantic Parsing via Span-level Supervised Attention (2021.naacl-main)
Copied to clipboard
Pengcheng Yin, Hao Fang, Graham Neubig, Adam Pauls, Emmanouil Antonios Platanios, Yu Su, Sam Thomson, Jacob Andreas
| Challenge: | Existing approaches to compositional generalization in semantic parsers focus on word-level alignments, but they focus on spans. |
| Approach: | They propose a span-level supervised attention loss that improves compositional generalization in semantic parsers by focusing on spans. |
| Outcome: | The proposed method improves on three benchmarks of compositional generalization. |
Emotion Transfer with Enhanced Prototype for Unseen Emotion Recognition in Conversation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing research on emotion recognition in conversation does not reach a consensus on classification theories . despite this, there is no clear consensus on how to recognize previously unseen emotions in real-world applications. |
| Approach: | They propose a prototype-based emotion transfer framework that can be used in real-world applications. |
| Outcome: | The proposed framework shows promise but still faces key challenges in the field of emotion recognition in conversation. |
LenAtten: An Effective Length Controlling Unit For Text Summarization (2021.findings-acl)
Copied to clipboard
| Challenge: | Fixed length summarization (FLS) requires generating summaries with a preset number of characters or words. |
| Approach: | They propose a length control unit called LenAtten to break this trade-off by generating a short and coherent summary with the target length. |
| Outcome: | The proposed model improves controllability and ROGUE scores and generalizes well. |
Active Sentence Learning by Adversarial Uncertainty Sampling in Discrete Space (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Existing uncertainty sampling methods are time-consuming and can't be executed frequently. |
| Approach: | They propose adversarial uncertainty sampling in discrete space to find informative unlabeled text samples for annotation using adversarials. |
| Outcome: | The proposed approach outperforms baselines on effectiveness on five datasets. |
Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring (2026.acl-long)
Copied to clipboard
| Challenge: | Large Vision-Language Models (LVLMs) are vulnerable to a growing array of multimodal jailbreak attacks, necessitating a generalizable defense that is efficient for practical deployment. |
| Approach: | They propose a framework that uses a lightweight projection to separate benign and malicious inputs in safety-critical layers. |
| Outcome: | The proposed framework enables a simple yet powerful contrastive score that differentiates true malicious intent from mere distribution shift. |
SERE: Structural Example Retrieval for Enhancing LLMs in Event Causality Identification (2026.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have demonstrated strong performance across various NLP tasks, but their effectiveness in ECI remains limited due to biases in causal reasoning. |
| Approach: | They propose a structural example retrieval framework that leverages LLMs’ few-shot learning capabilities to help LLM models in ECI. |
| Outcome: | The proposed framework leverages LLMs’ few-shot learning capabilities to guide LLM models in causal reasoning, mitigating bias and improving accuracy. |
RMTBench: Benchmarking LLMs Through Multi-Turn User-Centric Role-Playing (2025.findings-emnlp)
Copied to clipboard
Hao Xiang, Tianyi Tang, Yang Su, Bowen Yu, An Yang, Fei Huang, Yichang Zhang, Yaojie Lu, Hongyu Lin, Xianpei Han, Jingren Zhou, Junyang Lin, Le Sun
| Challenge: | Existing benchmarks focus on character-centric approach and fail to reflect real-world applications. |
| Approach: | RMTBench is a user-centric bilingual role-playing benchmark featuring 80 diverse characters and over 8,000 dialogue rounds. |
| Outcome: | RMTBench features 80 diverse characters and over 8,000 dialogue rounds. |
Reliable Use of Lemmas via Eligibility Reasoning and Section-Aware Reinforcement Learning (2026.acl-short)
Copied to clipboard
| Challenge: | Recent large language models (LLMs) perform strongly on mathematical benchmarks but often import conclusions without validating assumptions. |
| Approach: | They propose a model that encodes a lemma specification and trains with reinforcement learning and section-aware loss masking to assign penalty to the section responsible for errors. |
| Outcome: | The proposed model performs well on benchmarks but often misapplyes lemmas . the model is able to encode the specification and train with reinforcement learning . |
A Large Scale Speech Sentiment Corpus (2020.lrec-1)
Copied to clipboard
| Challenge: | Existing corpus for sentiment analysis uses text inputs, but voice inputs are becoming more important as smart assistants and mobile voice control become more prevalent. |
| Approach: | They propose to extend the Switchboard-1 Telephone Speech Corpus by adding sentiment labels from 3 different human annotators for every transcript segment. |
| Outcome: | The proposed corpus contains 49500 labeled speech segments covering 140 hours of audio. |
Reasoning-Enhanced Domain-Adaptive Pretraining of Multimodal Large Language Models for Short Video Content Governance (2025.emnlp-industry)
Copied to clipboard
Zixuan Wang, Yu Sun, Hongwei Wang, Baoyu Jing, Xiang Shen, Xin Dong, Zhuolin Hao, Hongyu Xiong, Yang Song
| Challenge: | Existing approaches to identifying inappropriate content require extensive human-labeled data and lack cross-issue generalization. |
| Approach: | They propose a reasoning-enhanced multimodal large language model (MLLM) pretraining paradigm for unified inappropriate content detection. |
| Outcome: | The proposed model improves the MLLM's performance in both zero-shot and supervised fine-tuning settings and shows strong generalization capabilities to emergent, previously unseen issues. |
Context-Aware Non-Autoregressive Document-Level Translation with Sentence-Aligned Connectionist Temporal Classification (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing studies employ autoregressive translation (AT) methods to encode sentences . however, the AT methods struggle with error accumulation when the length of sentences increases. |
| Approach: | They propose a context-aware non-autoregressive framework with the sentence-aligned connectionist temporal classification loss for document-level neural machine translation. |
| Outcome: | The proposed framework achieves 46X speedup on three benchmarks compared to strong baselines. |
AgentThink: A Unified Framework for Tool-Augmented Chain-of-Thought Reasoning in Vision-Language Models for Autonomous Driving (2025.findings-emnlp)
Copied to clipboard
Kangan Qian, Sicong Jiang, Yang Zhong, Ziang Luo, Zilin Huang, Tianze Zhu, Kun Jiang, Mengmeng Yang, Zheng Fu, Jinyu Miao, Yining Shi, He Zhe Lim, Li Liu, Tianbao Zhou, Hongyi Wang, Huang Yu, Yifei Hu, Guang Li, Guang Chen, Hao Ye, Lijun Sun, Diange Yang
| Challenge: | Vision-Language Models struggle with hallucinations, inefficient reasoning, and limited real-world validation hinders accurate perception and robust step-by-step reasoning. |
| Approach: | AgentThink integrates Chain-of-Thought reasoning with dynamic, agent-style tool invocation for autonomous driving tasks. |
| Outcome: | Experiments on the DriveLMM-o1 benchmark show AgentThink significantly boosts overall reasoning scores by 53.91% and enhances answer accuracy by 33.54% . |
Reverse Modeling in Large Language Models (2025.naacl-short)
Copied to clipboard
| Challenge: | Using pre-trained LLMs with reversed text inputs can improve their performance across multiple languages. |
| Approach: | They propose a way to determine whether LLMs can understand reversed text inputs by reversing entire paragraphs or documents at the token level. |
| Outcome: | The proposed model can be used to improve understanding across multiple languages. |
LoRAPrune: Structured Pruning Meets Low-Rank Parameter-Efficient Fine-Tuning (2024.findings-acl)
Copied to clipboard
| Challenge: | Low-rank adaption (LoRA) is a low-level pruning method that can be expensive and slow to deploy. |
| Approach: | They propose a low-rank adaption pruning framework that provides an accurate structured pruned model in a memory-efficient manner. |
| Outcome: | The proposed pruning framework reduces perplexity and memory usage by 52.6% on LLaMA and T5 models while reducing memory usage. |
Analyzing LLMs’ Knowledge Boundary Cognition Across Languages Through the Lens of Internal Representations (2025.acl-long)
Copied to clipboard
| Challenge: | Understanding the knowledge boundaries of Large Language Models (LLMs) is crucial to prevent hallucination, but research on the knowledge boundary perceptions of LLMs has predominantly focused on English. |
| Approach: | They propose a training-free alignment method that effectively transfers knowledge boundary perception ability across languages, thereby helping reduce hallucination risk in low-resource languages. |
| Outcome: | The proposed method reduces hallucination risk in low-resource languages by fine-tuning on bilingual question pair translation. |
Clip-Tuning: Towards Derivative-free Prompt Learning with a Mixture of Rewards (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing work does not take full advantage of over-parameterized characteristics of large pre-trained language models. |
| Approach: | They propose a method that uses frozen "thinned" networks to obtain a mixture of rewards and advance the derivative-free prompt learning. |
| Outcome: | The proposed method outperforms previous gradient-free prompt learning methods and achieves parity with gradient-based counterparts on seven language understanding benchmarks under few-shot settings. |
A Bounding Box is Worth One Token - Interleaving Layout and Text in a Large Language Model for Document Understanding (2025.findings-acl)
Copied to clipboard
Jinghui Lu, Haiyang Yu, Yanjie Wang, Yongjie Ye, Jingqun Tang, Ziwei Yang, Binghong Wu, Qi Liu, Hao Feng, Han Wang, Hao Liu, Can Huang
| Challenge: | Existing methods for integrating spatial layouts with text have limitations . existing methods produce overly long text sequences or lack autoregressive traits of LLMs . |
| Approach: | They introduce Interleaving Layout and Text in a Large Language Model (LayTextLLM) they use OCR-derived text and spatial layouts to integrate with LLMs for document understanding . |
| Outcome: | The proposed model shows an increase in performance in KIE and VQA tasks. |
MARCH: Multi-Agent Reinforced Check for Hallucination (2026.acl-long)
Copied to clipboard
Zhuo Li, Yupeng Zhang, Pengyu Cheng, Jiajun Song, Mengyu Zhou, Hao Li, Shujie Hu, Yu Qin, null Erchao.zec, Xiaoxi Jiang, null Guanjunjiang
| Challenge: | Existing methods to detect hallucinations suffer from inherent confirmation bias, where the verifier inadvertently reproduces the errors of the original generation. |
| Approach: | They propose a framework that enforces rigorous factual alignment by leveraging deliberate *information asymmetry* by combining a pipeline of three specialized agents: a Solver, a Proposer, and a Checker. |
| Outcome: | Extensive experiments across hallucination benchmarks demonstrate that MARCH substantially reduces hallucinism rates. |
FewRel: A Large-Scale Supervised Few-Shot Relation Classification Dataset with State-of-the-Art Evaluation (D18-1)
Copied to clipboard
| Challenge: | Empirical results show that even the most competitive few-shot learning models struggle on this task, especially as compared with humans. |
| Approach: | They propose a Few-Shot Relation Classification Dataset consisting of 70, 000 sentences on 100 relations derived from Wikipedia and annotated by crowdworkers. |
| Outcome: | The proposed methods perform well on the most competitive few-shot learning models, especially as compared with humans. |
KeFVP: Knowledge-enhanced Financial Volatility Prediction (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Current studies ignore the role of financial metrics knowledge in earnings calls and little consideration is given to integrating text and price information. |
| Approach: | They propose to integrate financial metrics knowledge into text comprehension by knowledge-enhanced adaptive pre-training and effectively incorporating text and price information by introducing a conditional time series prediction module. |
| Outcome: | The proposed method outperforms state-of-the-art methods on three real-world datasets and is effective and reliable. |
On the Emotion Understanding of Synthesized Speech (2026.acl-long)
Copied to clipboard
Yuan Ge, Haishu Zhao, AoKai Hao, Junxiang Zhang, Bei Li, Xiaoqian Liu, Chenglong Wang, Jianjin Wang, Bingsen Zhou, Bingyu Liu, JingBo Zhu, Zhengtao Yu, Tong Xiao
| Challenge: | Existing models for emotion understanding do not capture fundamental features of synthesized speech. |
| Approach: | They evaluate emotion recognition models on synthesized speech using SER models and generative models. |
| Outcome: | The proposed model can't generalize to synthesized speech because of speech token prediction . generative models tend to infer emotion from textual semantics while ignoring paralinguistic cues. |
RePair: Automated Program Repair with Process-based Feedback (2024.findings-acl)
Copied to clipboard
| Challenge: | Commercial-scale language models (LMs) have taken APR to unprecedented levels, but they are limited by parameters and humans interact with them through explicit prompts. |
| Approach: | They propose a method that utilizes process supervision to improve program repair by allowing users to input feedback from compilers and test cases. |
| Outcome: | The proposed method outperforms large outcome-based generation methods and is inspired by strategies used in programming competitions. |
Continual Learning for Multilingual Neural Machine Translation via Dual Importance-based Model Division (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods focus on preventing catastrophic forgetting by making compromises between the original and new language pairs, leading to sub-optimal performance on both translation tasks. |
| Approach: | They propose a dual importance-based model division method to divide the model parameters into two parts and separate the translation of the original and new tasks. |
| Outcome: | The proposed method outperforms strong baselines under different incremental translation scenarios. |
Improved Pseudo Data for Machine Translation Quality Estimation with Constrained Beam Search (2023.emnlp-main)
Copied to clipboard
Xiang Geng, Yu Zhang, Zhejian Lai, Shuaijie She, Wei Zou, Shimin Tao, Hao Yang, Jiajun Chen, Shujian Huang
| Challenge: | evaluating the quality of machine translation outputs becomes increasingly essential with the rapid development of machine language (MT). |
| Approach: | They propose to generate pseudo data using the MT model with constrained beam search (CBSQE) they propose to preserve the reference parts with high MT probabilities as correct translations . |
| Outcome: | The proposed model outperforms strong baselines in both supervised and unsupervised settings. |
Teaching Small Language Models Reasoning through Counterfactual Distillation (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) have demonstrated remarkable performance in a wide range of downstream tasks. |
| Approach: | They propose a counterfactual distillation framework that leverages LLMs to generate high-quality counterfacts and utilizes multi-view CoT to enhance the diversity of reasoning samples. |
| Outcome: | The proposed framework enhances reasoning capabilities of large language models and is more robust to OOD data. |
SimPBL: A Multi-Agent Framework for Project-Based Learning (2026.acl-long)
Copied to clipboard
Daniel Zhang-Li, Joy Jia Yin Lim, Binglin Liu, Shangqing Tu, Zijun Yao, Hao Peng, Jifan Yu, Haoxuan Li, Zhanxin Hao, Ye He, Zekun Li, Jiangyi Wang, Lei Hou, Bin Xu, Xin Cong, Zhiyuan Liu, Huiqin Liu, Yu Zhang, Juanzi Li
| Challenge: | Existing LLMs provide partial assistance without modeling these roles, and overly comprehensive help can reduce learner autonomy. |
| Approach: | They propose a multi-agent framework with an orchestrator agent that provides adaptive scaffolding from interaction logs and collaborator agents that support project work through boundary-aware collaboration. |
| Outcome: | The proposed framework improves learner examination scores by 14% . it is based on a multi-agent framework with an orchestrator agent . |
Mixture-of-LoRAs: An Efficient Multitask Tuning Method for Large Language Models (2024.lrec-main)
Copied to clipboard
| Challenge: | Instruction Tuning has the potential to stimulate or enhance specific capabilities of large language models. |
| Approach: | They propose a mixture-of-LoRAs architecture which is a parameter-efficient tuning method designed for multi-task learning with LLMs. |
| Outcome: | The proposed method can be iteratively adapted to a new domain, enabling quick domain-specific adaptation. |
Rethinking Text-based Protein Understanding: Retrieval or LLM? (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent studies have focused on integrating protein-related knowledge into large language models through continued pretraining and multi-modal alignment. |
| Approach: | They propose a retrieval-enhanced method which significantly outperforms fine-tuned LLMs for protein-to-text generation and shows accuracy and efficiency in training-free scenarios. |
| Outcome: | The proposed method significantly outperforms fine-tuned LLMs for protein-to-text generation and shows accuracy and efficiency in training-free scenarios. |
ERNIE-M: Enhanced Multilingual Representation by Aligning Cross-lingual Semantics with Monolingual Corpora (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for pretraining cross-lingual models are limited in their size due to the limited amount of parallel corpora. |
| Approach: | They propose a method that encourages the model to align multiple languages with monolingual corpora to overcome the constraint of the parallel corpus size. |
| Outcome: | The proposed method outperforms existing cross-lingual models and delivers new state-of-the-art results in various cross-linguistic downstream tasks. |
C2PO: Diagnosing and Disentangling Bias Shortcuts in LLMs (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to solve large language models address stereotypical and structural biases in isolation . however, prior paradigms address these in isolation, often at the expense of exacerbating the other . |
| Approach: | They propose a framework to tackle latent spurious feature correlations within input that drive erroneous reasoning shortcuts. |
| Outcome: | The proposed framework mitigates stereotypical and structural biases while preserving robust general reasoning capabilities. |
Adaptive Structure Induction for Aspect-based Sentiment Analysis with Spectral Perspective (2023.findings-emnlp)
Copied to clipboard
| Challenge: | incorporating structure information can enhance the performance of aspect-based sentiment analysis. |
| Approach: | They propose to use pre-trained language models to induct latent structures from a spectrum perspective. |
| Outcome: | The proposed model shortens Aspects-sentiment Distance and improves structure induction ability. |
Imagination and Contemplation: A Balanced Framework for Semantic-Augmented Multimodal Machine Translation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Multimodal Machine Translation (MMT) is effective in resolving linguistic ambiguities, but visual information often introduces redundancy or noise, potentially impairing translation quality. |
| Approach: | They propose a semantic-augmented framework that integrates "Imagination" and "Contemplation" they first generate synthetic images from source text and align them with authentic images via an optimal transport loss . |
| Outcome: | The proposed framework outperforms baselines on translation datasets with visually ambiguous or weakly correlated content. |
MdEval: Massively Multilingual Code Debugging (2026.findings-acl)
Copied to clipboard
Shukai Liu, Linzheng Chai, Jian Yang, Jiajun Shi, He Zhu, Liran Wang, Jin Ke, Wei Zhang, Hualei Zhu, Shuyue Guo, Tao Sun, Jiaheng Liu, Yunlong Duan, Yu Hao, Liqun Yang, Guanglin Niu, Ge Zhang, Zhoujun Li
| Challenge: | Existing benchmarks primarily focus on Python and are limited in terms of language diversity. |
| Approach: | They propose a multilingual debugging benchmark that includes 3.9K test samples of 20 programming languages and introduces the debug instruction corpora MdEval-Instruct by injecting bugs into the correct multilingual queries and solutions. |
| Outcome: | The proposed benchmark includes 3.9K test samples of 20 programming languages and covers the automated program repair task, bug localization task, and bug identification task. |
COSY: COunterfactual SYntax for Cross-Lingual Understanding (2021.acl-long)
Copied to clipboard
| Challenge: | Pre-trained multilingual language models suffer from a large performance gap between source and target languages . e.g., multilingual-BERT models are widely used in cross-lingual tasks . |
| Approach: | They propose a language-agnostic approach to integrate universal syntax into language models . they use SYntax-aware networks and a COunterfactual training method . |
| Outcome: | The proposed model achieves state-of-the-art performance on natural language inference and question answering without auxiliary training data. |
Knowing False Negatives: An Adversarial Training Method for Distantly Supervised Relation Extraction (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for relation extraction ignore the incompleteness of existing knowledge bases . current methods are too weak and cause noises when training and testing are not based on training data. |
| Approach: | They propose a method to automatically align unstructured text with relation instances in a knowledge base . they use heuristics to leverage the memory mechanism of deep neural networks to find out possible FN samples . |
| Outcome: | Experiments on two wildly-used benchmark datasets show the effectiveness of the proposed method. |
Domain Incremental Lifelong Learning in an Open World (2023.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to lifelong learning (LL) models require access to task identities in the testing phase or cannot handle samples from unseen tasks. |
| Approach: | They propose a dynamic architecture-based lifelong learning model that tries to learn a sequence of tasks with a prompt-enhanced language model. |
| Outcome: | The proposed model outperforms state-of-the-art models in handling unseen tasks and focuses on task-level prompts to capture knowledge from different granularities. |
Diversified Multiple Instance Learning for Document-Level Multi-Aspect Sentiment Classification (2020.emnlp-main)
Copied to clipboard
| Challenge: | Experimental results show that D-MILN outperforms recent weakly-supervised baselines . document-level multi-aspect sentiment classification requires a lot of manual aspect-level annotations - which is time-consuming and laborious . |
| Approach: | They propose a novel Diversified Multiple Instance Learning Network to achieve DMSC with only document-level weak supervision. |
| Outcome: | The proposed method outperforms weakly-supervised baselines on TripAdvisor and BeerAdvocate datasets. |
Translate-Train Embracing Translationese Artifacts (2022.acl-short)
Copied to clipboard
| Challenge: | Existing approaches to train multilingual tasks are based on translationese and translatetrain. |
| Approach: | They propose to use translationese to mitigate the gap between the source and target languages to train the translator. |
| Outcome: | The proposed method outperforms baselines on the multilingual QA dataset TyDiQA. |
SWEET - Weakly Supervised Person Name Extraction for Fighting Human Trafficking (2023.findings-emnlp)
Copied to clipboard
| Challenge: | SWEET is a weak supervision pipeline for extracting person names from noisy escort ads . it does not require any human annotators and labeling, which is incredibly important . |
| Approach: | They propose a weak supervision pipeline SWEET: Supervise Weakly for Entity Extraction to fight Trafficking for extracting person names from noisy escort ads. |
| Outcome: | The proposed weak supervision pipeline outperforms the previous method by 9% on domain data and generalizes to common benchmark datasets. |
Beyond Facts: Evaluating Intent Hallucination in Large Language Models (2025.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) produce unsatisfactory results when faced with complex queries containing multiple conditions. |
| Approach: | They propose a benchmark for intent hallucination that covers 20,068 problems and an automatic LLM generation evaluation metric for detecting intent hallucinosis. |
| Outcome: | The proposed benchmark covers query-only and retrieval-augmented generation (RAG) setups with varying topics and difficulty. |
Reinforced Counterfactual Data Augmentation for Dual Sentiment Classification (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to improve generalization ability by augmenting training data with synonymous examples or adding random noises to word embeddings cannot address spurious association problem. |
| Approach: | They propose an end-to-end reinforcement learning framework which jointly performs counterfactual data generation and dual sentiment classification. |
| Outcome: | The proposed framework outperforms strong data augmentation baselines on several benchmark sentiment classification datasets. |
Multi-View Incongruity Learning for Multimodal Sarcasm Detection (2025.coling-main)
Copied to clipboard
Diandian Guo, Cong Cao, Fangfang Yuan, Yanbing Liu, Guangjie Zeng, Xiaoyan Yu, Hao Peng, Philip S. Yu
| Challenge: | Existing methods for multimodal sarcasm detection rely on spurious correlations, demonstrating poor generalizability beyond training environments. |
| Approach: | They propose a method that integrates multimodal incongruities via contrastive learning for multimodal sarcasm detection by using three views to drive multi-view learning. |
| Outcome: | The proposed method outperforms existing methods on benchmark datasets and shows that it is more generalizable than existing methods. |
Chain-of-Skills: A Configurable Model for Open-Domain Question Answering (2023.acl-long)
Copied to clipboard
| Challenge: | Using customized retrieval models, model transferability and scalability are limited. |
| Approach: | They propose a modular retrieval model where individual modules correspond to key skills that can be reused across datasets. |
| Outcome: | The proposed model outperforms self-supervised retrievers in zero-shot evaluations and achieves state-of-the-art fine-tuned retrieval performance on NQ, HotpotQA and OTT-QA. |
Learning Logic Rules for Document-Level Relation Extraction (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing models for document-level relation extraction relied on implicitly powerful representations, which makes the model less transparent. |
| Approach: | They propose a probabilistic model for document-level relation extraction by learning logic rules. |
| Outcome: | The proposed model outperforms baseline models in relation performance and logical consistency. |
AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents (2025.acl-long)
Copied to clipboard
Yifan Xu, Xiao Liu, Xueqiao Sun, Siyi Cheng, Hao Yu, Hanyu Lai, Shudan Zhang, Dan Zhang, Jie Tang, Yuxiao Dong
| Challenge: | Existing studies on Android agents lack systematic research on open-source and closed-source models. |
| Approach: | They propose a framework for Android agents that includes an operation environment and a reproducible benchmark. |
| Outcome: | The proposed framework lifts the success rate of open-source LLMs and LMMs from 4.59% to 21.50% for LLM and 1.93% to 13.28% for LMM. |
PRA-RAG: Provably Robust Aggregation in Retrieval-Augmented Generation against Retrieval Corruption (2026.findings-acl)
Copied to clipboard
Xue Tan, Yi Zheng, Chang Huo, Yunruo Zhang, Yu Liu, Hao Luan, Zhuyang Yu, Jun Dai, Xiaoyan Sun, Ping Chen
| Challenge: | Existing defense mechanisms lack theoretical robustness guarantees and perform unreliably when the LLM has limited knowledge of the retrieved content. |
| Approach: | They propose a provably robust retrieval aggregation algorithm designed to defend against poisoning attacks on retrieved texts. |
| Outcome: | Experiments show that PRA-RAG reduces the attack success rate to as low as 1% while maintaining an accuracy of 71%, significantly outperforming representative state-of-the-art (SOTA) methods. |
C-ReD: A Comprehensive Chinese Benchmark for AI-Generated Text Detection Derived from Real-World Prompts (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent efforts to develop algorithms for large language models (LLMs) have limited model diversity and data homogeneity in the Chinese corpora. |
| Approach: | They propose a Chinese Real-prompt AI-generated text Detection benchmark that can be generalized to unseen LLMs and external Chinese datasets. |
| Outcome: | The proposed benchmarks address critical gaps in model diversity, domain coverage, and prompt realism that have limited prior Chinese detection benchmarks. |
Self-Steering Optimization: Autonomous Preference Optimization for Large Language Models (2025.findings-acl)
Copied to clipboard
Hao Xiang, Bowen Yu, Hongyu Lin, Keming Lu, Yaojie Lu, Xianpei Han, Ben He, Le Sun, Jingren Zhou, Junyang Lin
| Challenge: | Prior research focused on developing data generation methods, while insufficient attention has been paid to quality control mechanisms and often produces inaccurate and unhelpful data. |
| Approach: | They propose an algorithm that automatically generates high-quality preference data, eliminating manual annotation requirements. |
| Outcome: | The proposed algorithm outperforms baselines in human preference alignment and reward optimization. |
Self-Reinforcing Controllable Synthesis of Rare Relational Data via Bayesian Calibration (2026.findings-acl)
Copied to clipboard
Chongsheng Zhang, Hao Wang, Zelong Yu, Esteban Garces Arias, Julian Rodemann, Zhanshuo Zhang, Qilong Li, Gaojuan Fan, Krikamol Muandet, Christian Heumann
| Challenge: | Existing approaches to synthesis of relational/structured tabular data lack effective feedback mechanism to optimize quality of generated data. |
| Approach: | They propose a relational data generator with dynamic guidance framework that uses chain-of-thought steps to generate tabular data for enhancing downstream imbalanced classification performance. |
| Outcome: | The proposed framework outperforms existing approaches in both data fidelity and downstream imbalanced classification performance on real and synthetic datasets. |
AJ-Bench: Benchmarking Agent-as-a-Judge for Environment-Aware Evaluation (2026.findings-acl)
Copied to clipboard
Wentao Shi, Yu Wang, Yuyang Zhao, Yuxin Chen, Fuli Feng, Xueyuan Hao, Xi Su, Qi GU, Hui Su, Xunliang Cai, Xiangnan He
| Challenge: | Existing approaches to verify agent behaviors in complex environments rely on rule-based verifiers or LLM-as-a-Judge models. |
| Approach: | They propose a benchmark to evaluate Agent-as-a-Judge across three domains . the benchmark covers search, data systems, and graphical user interfaces - with 155 tasks and 516 trajectories . |
| Outcome: | The proposed benchmark outperforms existing benchmarks in search, data systems, and GUI domains while revealing open challenges in agent-based verification. |
ACEBench: A Comprehensive Evaluation of LLM Tool Usage (2025.findings-emnlp)
Copied to clipboard
Chen Chen, Xinlong Hao, Weiwen Liu, Xu Huang, Xingshan Zeng, Shuai Yu, Dexun Li, Yuefeng Huang, Xiangcheng Liu, Wang Xinzhi, Wu Liu
| Challenge: | Existing benchmarks for evaluating LLMs’ tool usage face several limitations: limited evaluation scenarios, lacking assessments in real multi-turn dialogue contexts; narrow evaluation dimensions, with insufficient detailed assessments of how LLM use tools; and reliance on LLM or real API executions for evaluation, which introduces significant overhead. |
| Approach: | ACEBench is a benchmark for evaluating tool usage in Large Language Models . it categorizes data into three primary types based on evaluation methodology: Normal, Special, and Agent. |
| Outcome: | ACEBench categorizes data into three primary types based on evaluation methodology: Normal, Special, and Agent. |
Understanding Deep Learning Performance through an Examination of Test Set Difficulty: A Psychometric Case Study (D18-1)
Copied to clipboard
| Challenge: | Existing methods to evaluate deep learning models that are not considered for test set accuracy are difficult to interpret. |
| Approach: | They examine the impact of a test set question’s difficulty to determine if there is a relationship between difficulty and performance. |
| Outcome: | The proposed model can learn examples of varying difficulty at different rates if it does well on hard examples and poor on easy items because a dataset is all easy, but has "solved" anything? |
LLMs know their vulnerabilities: Uncover Safety Gaps through Natural Distribution Shifts (2025.acl-long)
Copied to clipboard
Qibing Ren, Hao Li, Dongrui Liu, Zhanxu Xie, Xiaoya Lu, Yu Qiao, Lei Sha, Junchi Yan, Lizhuang Ma, Jing Shao
| Challenge: | Current safety training focuses on teaching models to reject harmful queries, but recent research shows that adversarial attacks or jailbreak methods bypass these safety mechanisms. |
| Approach: | They propose to use a new attack method to craft multi-turn toxic prompts that gradually lead LLMs to reveal unsafe content. |
| Outcome: | The proposed method outperforms existing methods in diversity, effectiveness, and efficiency across aligned LLMs. |
Optimizing Instruction Synthesis: Effective Exploration of Evolutionary Space with Tree Search (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Extensive research has highlighted the quality of instruction data is essential for the success of this alignment. |
| Approach: | They propose a framework for iteratively improving existing instruction data by using Monte Carlo tree search to find suitable prompts that align the language model to effectively learn multiple skills. |
| Outcome: | The proposed framework improves the evaluation scores of seed instruction data, raising the average evaluation scores from 2.19 to 3.81. |
Dynamically Fused Graph Network for Multi-hop Reasoning (P19-1)
Copied to clipboard
| Challenge: | Text-based question answering (TBQA) has been studied extensively in recent years. |
| Approach: | They propose a Dynamically Fused Graph Network to answer questions requiring multiple scattered evidence and reasoning over them. |
| Outcome: | The proposed method achieves competitive results on a public TBQA dataset and produces interpretable reasoning chains. |
HTCInfoMax: A Global Model for Hierarchical Text Classification via Information Maximization (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing models for hierarchical text classification do not consider statistical constraint on label representations learned by structure encoder. |
| Approach: | They propose a new hierarchical text classification model called HTCInfoMax which incorporates two modules to improve the model's representations. |
| Outcome: | The proposed model can model the interaction between each text sample and its ground truth labels explicitly which filters out irrelevant information. |
Divide and Conquer: Legal Concept-guided Criminal Court View Generation (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for creating rationales for criminal cases do not pay enough attention to the important legal concepts. |
| Approach: | They propose a legal concept-guided court view generation framework that generates rationales based on predicted legal concepts . they first divide the court view into sub-views, then employ a solver and verifier to generate and select rationale. |
| Outcome: | The proposed model generates coherent and coherent court views on a real-world criminal case dataset. |
LPO: Towards Accurate GUI Agent Interaction via Location Preference Optimization (2026.findings-acl)
Copied to clipboard
Jiaqi Tang, Yu Xia, Yi-Feng Wu, Yuwei Hu, Chen Yuhui, Qing-Guo Chen, Xiaogang Xu, Xiangyu Wu, Hao LU, Yanqing Ma, Shiyin Lu, Qifeng Chen
| Challenge: | Existing strategies for spatial localization are limited due to their limited capacity to perceive positional data. |
| Approach: | They propose a location-based approach that leverages locational data to optimize interaction preferences. |
| Outcome: | The proposed approach achieves SOTA results across offline benchmarks and real-world evaluations. |
Lexicon-Based Graph Convolutional Network for Chinese Word Segmentation (2021.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for Chinese word segmentation have high performance on benchmarks but are limited by the small-scale annotated corpus. |
| Approach: | They propose a framework that incorporates a lexicon-based graph convolutional network into the Transformer encoder to improve Chinese word segmentation (CWS) Chinese word is an essential and pre-processing step for many downstream NLP tasks. |
| Outcome: | The proposed framework captures the information of candidate words and improves performance on benchmarks and datasets. |
Scaling Data Diversity for Fine-Tuning Language Models in Human Alignment (2024.lrec-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) can reveal toxic or offensive content inadvertently or intentionally. |
| Approach: | They propose to control the diversity of both sides according to the number of samples for fine-tuning, which can directly reflect their impact. |
| Outcome: | The proposed approach improves the performance of large language models after fine-tuning. |
Training Verifier to Assessing Complex Real-World Tool-Use Trajectories (2026.findings-acl)
Copied to clipboard
Linzhuang Sun, Mingyang Chen, Hao Liang, Tianpeng Li, Zhou Yijie, Chenzheng Zhu, Tianyu Guo, Huanyao Zhang, Jingxuan Wei, Bihui Yu, Fan Yang, Wentao Zhang
| Challenge: | Existing methods for training effective AI agents often resort to synthetic data generation. |
| Approach: | They propose a plug-and-play framework for data quality control in tool-use scenarios . they construct a tool-verify dataset and release a benchmark to assess its performance . |
| Outcome: | The proposed framework surpasses Qwen2.5-72B-Instruct on Tool-V-Bench and the previous APIGen-MT dataset. |
Explain the Synth: Interpretable Evaluation of LLM Data Synthesis (2026.acl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) are increasingly used to generate tabular data. |
| Approach: | They propose a framework that uses a rule-based model as a shared explanatory language to examine the explanation of real versus synthetic data. |
| Outcome: | The proposed framework compares the explanatory structure induced by real versus synthetic data. |
FineReason: Evaluating and Improving LLMs’ Deliberate Reasoning through Reflective Puzzle Solving (2025.acl-long)
Copied to clipboard
Guizhen Chen, Weiwen Xu, Hao Zhang, Hou Pong Chan, Chaoqun Liu, Lidong Bing, Deli Zhao, Anh Tuan Luu, Yu Rong
| Challenge: | Recent advances in large language models (LLMs) highlight an important shift from the “System 1” way of quick reactions to the “system 2” style of reflection-and-correction problem solving. |
| Approach: | They propose a logic-puzzle benchmark for systematic evaluation of large language models' reasoning capabilities that decomposes each puzzle into atomic steps. |
| Outcome: | The proposed model improves on state checking and state transition tasks and demonstrates gains in reasoning by up to 5.1%. |
WalledEval: A Comprehensive Safety Evaluation Toolkit for Large Language Models (2024.emnlp-demo)
Copied to clipboard
Prannaya Gupta, Le Yau, Hao Low, I-Shiang Lee, Hugo Lim, Yu Teoh, Koh Hng, Dar Liew, Rishabh Bhardwaj, Rajat Bhardwaj, Soujanya Poria
| Challenge: | Potential harms include training data leakage, biases in responses and decision-making, and unauthorized use for purposes such as terrorism and the generation of sexually explicit content. |
| Approach: | WalledEval is a comprehensive AI safety testing toolkit designed to evaluate large language models. |
| Outcome: | The framework supports both LLM and judge benchmarking and incorporates custom mutators to test safety against various text-style mutations such as future tense and paraphrasing. |
Text Style Transfer Back-Translation (2023.acl-long)
Copied to clipboard
Daimeng Wei, Zhanglin Wu, Hengchao Shang, Zongyao Li, Minghan Wang, Jiaxin Guo, Xiaoyu Chen, Zhengzhe Yu, Hao Yang
| Challenge: | Current methods require large amount of bilingual training data, which is challenging and sometimes impossible task. |
| Approach: | They propose a method to modify the style of inputs by modifying the source side of BT data. |
| Outcome: | The proposed method significantly improves translation quality against popular BT benchmarks on high-resource and low-resourced language pairs. |
Learning Latent Parameters without Human Response Patterns: Item Response Theory with Artificial Crowds (D19-1)
Copied to clipboard
| Challenge: | Incorporating Item Response Theory (IRT) into NLP tasks can provide valuable information about model performance and behavior. |
| Approach: | They propose to use IRT models generated from artificial crowds of DNNs to learn IRT. |
| Outcome: | The proposed model learning method outperforms baseline methods for two NLP tasks. |
TLUE: A Tibetan Language Understanding Evaluation Benchmark (2025.emnlp-main)
Copied to clipboard
Fan Gao, Cheng Huang, Yutong Liu, Nyima Tashi, Xiangxiang Wang, Thupten Tsering, Ban Ma-bao, Renzeng Duojie, Gadeng Luosang, Rinchen Dongrub, Dorje Tashi, Xiao Feng Cd, Yongbin Yu, Hao Wang
| Challenge: | Low-resource languages, like Tibetan, remain underrepresented in large language models' evaluations. |
| Approach: | They propose a Tibetan Language Understanding Evaluation Benchmark to assess LLMs' proficiency in Tibetan . they use a multi-task understanding benchmark and a safety benchmark to evaluate models . |
| Outcome: | The proposed benchmark shows that most large language models perform below the random baseline, especially in Tibetan language processing. |
Glancing Transformer for Non-Autoregressive Neural Machine Translation (2021.acl-long)
Copied to clipboard
| Challenge: | Existing non-autoregressive neural machine translation methods are either inferior to Transformer or require multiple decoding passes, leading to reduced speedup. |
| Approach: | They propose a Glancing Language Model (GLM) for single-pass parallel generation models and Glancing Transformer (GLAT) with only single- pass decoding, GLAT is able to generate high-quality translation with 8-15 speedup. |
| Outcome: | The proposed model outperforms all previous non-autoregressive methods on multiple language directions and is nearly comparable to Transformer. |
Rethinking Entropy Interventions in RLVR: An Entropy Change Perspective (2026.acl-long)
Copied to clipboard
Zhezheng Hao, Hong Wang, Haoyang Liu, Jian Luo, Jiarui Yu, Hande Dong, Qiang Lin, Can Wang, Jiawei Chen
| Challenge: | Large Language Models (LLMs) have remarkable reasoning capabilities in complex tasks such as mathematics and coding. |
| Approach: | They propose an entropy-modulation method that adaptively reweighs tokens based on theoretically-estimated entropic variations. |
| Outcome: | The proposed method outperforms state-of-the-art methods in six mathematical reasoning and three coding benchmarks. |
LEPO: Latent Reasoning Policy Optimization for Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing latent reasoning methods that use chain of thought (CoT) are limited to selecting one discrete token at each reasoning step, which potentially induces information loss. |
| Approach: | They propose a framework that injects controllable stochasticity into latent reasoning via Gumbel-Softmax, restoring LLMs' exploratory capacity and enhancing their compatibility with Reinforcement Learning (RL). |
| Outcome: | The proposed framework preserves richer information for more comprehensive reasoning and is compatible with Reinforcement Learning (RL). |
Simulating Classroom Education with LLM-Empowered Agents (2025.naacl-long)
Copied to clipboard
Zheyuan Zhang, Daniel Zhang-Li, Jifan Yu, Linlu Gong, Jinchang Zhou, Zhanxin Hao, Jianxiao Jiang, Jie Cao, Huiqin Liu, Zhiyuan Liu, Lei Hou, Juanzi Li
| Challenge: | Initial studies have focused on task-specific, independent LLM-empowered agents, but the potential of LLMs within a multi-agent collaborative framework for classroom simulation with real user participation remains unexplored. |
| Approach: | They propose a multi-agent classroom simulation teaching framework that recognizes representative class roles and introduces a novel class control mechanism for automatic classroom teaching. |
| Outcome: | The proposed framework can simulate dynamic learning environment for users with active teacher-student and student-studente interactions. |
GeoPQA: Bridging the Visual Perception Gap in MLLMs for Geometric Reasoning (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Recent advances in reinforcement learning (RL) have enhanced the reasoning abilities of large language models, but the impact on multimodal LLMs is limited. |
| Approach: | They propose a two-stage RL framework that enhances visual perception and fosters reasoning capabilities. |
| Outcome: | The proposed framework improves geometric reasoning by 9.7% and problem-solving by 9.1% compared to direct reasoning training approach. |
RAST: Domain-Robust Dialogue Rewriting as Sequence Tagging (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing models for dialogue rewriting suffer from the robustness issue, i.e., performances drop dramatically when testing on a different dataset. |
| Approach: | They propose a sequence-tagging-based approach that reduces the search space while preserving the core of the task. |
| Outcome: | The proposed model significantly reduces the search space while still covering the core of the task. |
When More Data Hurts: A Troubling Quirk in Developing Broad-Coverage Natural Language Understanding Systems (2022.emnlp-main)
Copied to clipboard
Elias Stengel-Eskin, Emmanouil Antonios Platanios, Adam Pauls, Sam Thomson, Hao Fang, Benjamin Van Durme, Jason Eisner, Yu Su
| Challenge: | In natural language understanding systems, users’ evolving needs necessitate the addition of new features over time, indexed by new symbols added to the meaning representation space. |
| Approach: | They propose to use a small set of new symbols to build broad-coverage NLU systems. |
| Outcome: | The proposed model is based on two prototypical NLU tasks: intent recognition and semantic parsing. |
latent-GLAT: Glancing at Latent Variables for Parallel Text Generation (2022.acl-long)
Copied to clipboard
| Challenge: | Recent advances in text generation have limited applications due to multimodality problem. |
| Approach: | They propose a method which uses latent variables to capture word categorical information and invoke an advanced curriculum learning technique to overcome multi-modality problem. |
| Outcome: | The proposed method outperforms strong baselines without an autoregressive model, which further broadens the application scenarios of the parallel decoding paradigm. |
LLMs in the Imaginarium: Tool Learning through Simulated Trial and Error (2024.acl-long)
Copied to clipboard
| Challenge: | Existing work on tool-augmented LLMs focuses on the broad coverage of tools and the flexibility of adding new tools. |
| Approach: | They propose a biologically inspired method for tool-augmented LLMs that orchestrates three key mechanisms for successful tool use behaviors in the biological system: trial and error, imagination, and memory. |
| Outcome: | The proposed method improves tool learning for LLMs under both in-context learning and fine-tuning settings, bringing a boost of 46.7% to Mistral-Instruct-7B and outperforms GPT-4. |
Middleware for LLMs: Tools Are Instrumental for Language Agents in Complex Environments (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) are generalist agents capable of operating within complex environments. |
| Approach: | They propose a class of tools that can serve as a middleware layer shielding LLMs from environmental complexity. |
| Outcome: | The proposed tool can shield the LLM from environmental complexity in two representative complex environments. |
Dynamic Collaboration of Multi-Language Models based on Minimal Complete Semantic Units (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to enhance reasoning capabilities of language models are expensive and often lack the ability to perform complex reasoning tasks. |
| Approach: | They propose a token-level multi-model collaboration strategy to enhance reasoning capabilities in language models by selecting the optimal tokens from the next token distributions. |
| Outcome: | The proposed method is superior to existing methods and will be released soon. |
AceGPT, Localizing Large Language Models in Arabic (2024.naacl-long)
Copied to clipboard
Huang Huang, Fei Yu, Jianqing Zhu, Xuening Sun, Hao Cheng, Song Dingjie, Zhihong Chen, Mosen Alharthi, Bang An, Juncai He, Ziche Liu, Junying Chen, Jianquan Li, Benyou Wang, Lian Zhang, Ruoyu Sun, Xiang Wan, Haizhou Li, Jinchao Xu
| Challenge: | Significant concerns emerge when addressing cultural sensitivity and local values. |
| Approach: | They propose a localized Large Language Model (LLM) specifically for Arabic, a language imbued with unique cultural characteristics inadequately addressed by current mainstream models. |
| Outcome: | The proposed model sets the state-of-the-art standard for open Arabic LLMs across various benchmarks. |
Posterior Differential Regularization with f-divergence for Improving Model Robustness (2021.naacl-main)
Copied to clipboard
| Challenge: | Recent studies show that pre-trained models suffer catastrophic degradation in out-of-domain generalization to datasets with domain shift or adversarial scenarios. |
| Approach: | They propose to regularize the posterior difference between clean and noisy inputs by using a Jacobian regularization framework and a virtual adversarial training framework. |
| Outcome: | The proposed framework can improve model robustness in fully supervised and semi-supervised settings. |
QAEncoder: Towards Aligned Representation Learning in Question Answering Systems (2025.acl-long)
Copied to clipboard
Zhengren Wang, Qinhan Yu, Shida Wei, Zhiyu Li, Feiyu Xiong, Xiaoxing Wang, Simin Niu, Hao Liang, Wentao Zhang
| Challenge: | Modern QA systems entail retrieval-augmented generation (RAG) for accurate and trustworthy responses, but the inherent gap between user queries and relevant documents hinders precise matching. |
| Approach: | They propose a retrieval-augmented generation (RAG)-based approach to bridge this gap by attaching document fingerprints to the embedding to estimate the expectation of potential queries. |
| Outcome: | Experiments across diverse datasets, languages, and embedding models confirm the proposed solution is simple-yet-effective with zero additional index storage, retrieval latency, training costs, or catastrophic forgetting and hallucination issues. |
ERNIE-Code: Beyond English-Centric Cross-lingual Pretraining for Programming Languages (2023.findings-acl)
Copied to clipboard
| Challenge: | ERNIE-Code is a unified pre-trained language model for 116 NLs and 6 PLs. |
| Approach: | They propose a unified pre-trained language model for 116 NLs and 6 PLs . they employ span-corruption language modeling that learns patterns from monolingual NL or PL . |
| Outcome: | The proposed model outperforms previous multilingual models for NL or NL across end tasks. |
LM-Interview: An Easy-to-use Smart Interviewer System via Knowledge-guided Language Model Exploitation (2024.emnlp-demo)
Copied to clipboard
| Challenge: | Semi-structured interviews are a crucial method of data acquisition in qualitative research. |
| Approach: | They propose a semi-structured interview system that automates interview preparation, analysis and control by interviewers. |
| Outcome: | Experimental results show that LM-Interview performs comparable to human interviewers . the system can be used to analyze semi-structured interviews without interviewers' involvement . |
CoreCodeBench: Decoupling Code Intelligence via Fine-Grained Repository-Level Tasks (2026.acl-long)
Copied to clipboard
Lingyue Fu, Hao Guan, Bolun Zhang, Haowei Yuan, Yaoming Zhu, Lin Qiu, ZongYu Wang, Xuezhi Cao, Xunliang Cai, Weiwen Liu, Weinan Zhang, Yong Yu
| Challenge: | Existing large language models for software engineering rely on coarse-grained pass rates obscuring specific cognitive bottlenecks. |
| Approach: | They propose a repository-level benchmark that dissects coding capabilities through atomized tasks. |
| Outcome: | The proposed framework achieves a 78.55% validity yield, surpassing the 31.7% retention rate of SWE-bench-Verified. |
MM-Verify: Enhancing Multimodal Reasoning with Chain-of-Thought Verification (2025.acl-long)
Copied to clipboard
| Challenge: | MM-Verifier and MM Reasoner are a powerful multimodal reasoning model . large language models (LLMs) have demonstrated exceptional performance across tasks spanning myriad domains. |
| Approach: | They propose a method which combines tree search and verification to generate high-quality chain-of-thought data. |
| Outcome: | The proposed method outperforms all larger models on the MathCheck, MathVista, and MathVerse benchmarks. |
Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models (2026.acl-long)
Copied to clipboard
Xin Cheng, Wangding Zeng, Damai Dai, Qinyu Chen, Bingxuan Wang, Zhenda Xie, Kezhao Huang, Xingkai Yu, Zhewen Hao, Han Zhang, Yu-Kun Li, Huishuai Zhang, Dongyan Zhao, Wenfeng Liang
| Challenge: | Mixture-of-Experts (MoE) scales capacity via conditional computation, but lacks knowledge lookup primitive. |
| Approach: | They propose a conditional memory instantiated via Deep Sparse Embedding (DSE) they propose 'u-shaped scaling law' that identifies optimal balance between MoE experts and DSE memory . |
| Outcome: | The proposed model outperforms an iso-parameter and isoFLOPs MoE baseline across knowledge and reasoning benchmarks and is infrastructure-efficient. |
AgenticRAGTracer: A Hop-Aware Benchmark for Diagnosing Multi-Step Retrieval Reasoning in Agentic RAG (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing benchmarks provide only final questions and answers, while lacking intermediate hop-level questions that gradually connect atomic questions to the final multi-hop query. |
| Approach: | They propose to build a multi-hop reasoning model that is primarily constructed automatically by large language models and designed to support step-by-step validation. |
| Outcome: | The proposed benchmark spans multiple domains, contains 1,305 data points, and has no overlap with existing mainstream benchmarks. |
Hierarchical Bi-Directional Self-Attention Networks for Paper Review Rating Recommendation (2020.coling-main)
Copied to clipboard
| Challenge: | Existing methods for review rating prediction ignore hierarchies among data . paper review rating predictions are important for improving paper review process . |
| Approach: | They propose a Hierarchical bi-directional self-attention Network framework for paper review rating prediction and recommendation . they leverage hierarchical structure of paper reviews with three levels of encoders . |
| Outcome: | The proposed approach can be used to make an effective decision-making tool for the academic paper review process. |
Neeko: Leveraging Dynamic LoRA for Efficient Multi-Character Role-Playing Agent (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have revolutionized open-domain dialogue agents but face challenges in multi-character role-playing (MCRP) scenarios. |
| Approach: | They propose a framework for efficient multi-character role-playing that employs a dynamic low-rank adapter strategy and distinct LoRA blocks for each character. |
| Outcome: | Neeko employs a dynamic low-rank adapter (LoRA) strategy, enabling it to adapt seamlessly to diverse characters. |
ERNIE-Gram: Pre-Training with Explicitly N-Gram Masked Language Modeling for Natural Language Understanding (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing methods to model coarse-grained linguistic information do not integrate coarse-gram information into pre-training. |
| Approach: | They propose an explicitly n-gram masking method to enhance integration of coarse-grained linguistic information into pre-training. |
| Outcome: | The proposed method outperforms existing models on English and Chinese text corpora and fine-tunes on 19 downstream tasks. |
AfriqueLLM: How Data Mixing and Model Architecture Impact Continued Pre-training for African Languages (2026.acl-long)
Copied to clipboard
Hao Yu, Tianyi Xu, Michael A. Hedderich, Wassim Hamidouche, Syed Waqas Zamir, David Ifeoluwa Adelani
| Challenge: | Continued pretraining (CPT) is a practical route to language adaptation, but improvements on demanding capabilities such as mathematical reasoning are limited. |
| Approach: | They propose to use CPT to adapt large language models to African languages . they use math, code, and synthetic translated data to analyze their models . |
| Outcome: | The proposed models improve on multilingual benchmarks and document-level translation. |
LVLM-Compress-Bench: Benchmarking the Broader Impact of Large Vision-Language Model Compression (2025.findings-naacl)
Copied to clipboard
Souvik Kundu, Anahita Bhiwandiwalla, Sungduk Yu, Phillip Howard, Tiep Le, Sharath Nittur Sridhar, David Cobbley, Hao Kang, Vasudev Lal
| Challenge: | LVLMs have been shown to perform well on simple uni-modal benchmarks, but their detailed study on multi-modal models is still lacking. |
| Approach: | They propose a framework to analyze the impact of compression on LVLMs on multi-modal input driven tasks. |
| Outcome: | The proposed framework analyzes the impact of compression on generative performance of large vision language models on multi-modal input driven tasks. |
INJONGO: A Multicultural Intent Detection and Slot-filling Dataset for 16 African Languages (2025.acl-long)
Copied to clipboard
Hao Yu, Jesujoba Oluwadara Alabi, Andiswa Bukula, Jian Yun Zhuang, En-Shiun Annie Lee, Tadesse Kebede Guge, Israel Abebe Azime, Happy Buzaaba, Blessing Kudzaishe Sibanda, Godson Koffi Kalipe, Jonathan Mukiibi, Salomon Kabongo Kabenamualu, Mmasibidi Setaka, Lolwethu Ndolela, Nkiruka Odu, Rooweither Mabuya, Shamsuddeen Hassan Muhammad, Salomey Osei, Sokhar Samb, Dietrich Klakow, David Ifeoluwa Adelani
| Challenge: | Slot-filling and intent detection tasks are well-established tasks in Conversational AI, but current benchmarks for these tasks rely on evaluations of low-resource languages and translations from English benchmarks. |
| Approach: | They propose to use a multilingual, open-source benchmark dataset for 16 African languages with utterances generated by native speakers across diverse domains. |
| Outcome: | The proposed dataset compares multilingual transformer models and prompting large language models (LLMs) with the English language. |
HopRAG: Multi-Hop Reasoning for Logic-Aware Retrieval-Augmented Generation (2025.findings-acl)
Copied to clipboard
| Challenge: | Traditional retrieval systems focus on lexical or semantic similarity rather than logical relevance. |
| Approach: | They propose a new RAG framework that augments retrieval with logical reasoning . hopRAG uses a retrieve-reason-prune mechanism to explore multi-hop neighbors . |
| Outcome: | The proposed framework outperforms conventional retrieval systems and state-of-the-art benchmarks on multi-hop QA tasks. |
Learning to Navigate Unseen Environments: Back Translation with Environmental Dropout (N19-1)
Copied to clipboard
| Challenge: | Existing approaches perform significantly worse in unseen environments compared to seen ones. |
| Approach: | They propose to use a ‘environmental dropout’ method to generate unseen triplets to generate new paths and instructions to generalize the agent. |
| Outcome: | The proposed agent outperforms the state-of-the-art approaches on the private unseen test set and is ranked top on the leaderboard. |
Defenses Against Prompt Attacks Learn Surface Heuristics (2026.acl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) are increasingly deployed in security-sensitive applications . recent defenses rely on supervised fine-tuning with benign and malicious labels . position bias arises when benign content placed later in a prompt is rejected at much higher rates . |
| Approach: | They analyze three recurring shortcut behaviors induced by supervised fine-tuning . position bias arises when benign content placed later in a prompt is rejected . token trigger bias occurs when strings common in attack data raise rejection probability . |
| Outcome: | The proposed model overrides intended logic when adversarial instructions appear . the proposed model has low rejection rates but narrow correlations in defense data . |
Revisiting Audio-language Pretraining for Learning General-purpose Audio Representation (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for learning general-purpose audio representations are limited in scope and coverage of audio attributes. |
| Approach: | They propose to use a 10.7M caption dataset to compare ALP with captioning . they find that contrastive learning yields competitive, transferable representations . |
| Outcome: | The proposed model yields competitive, transferable representations, while captioning exhibits better scalability. |
Generating Visual Spatial Description via Holistic 3D Scene Understanding (2023.acl-long)
Copied to clipboard
| Challenge: | Existing VSD work focuses on skewed spatial understanding of target objects . Existing work merely models the 2D geometrical vision features . |
| Approach: | They propose to incorporate 3D scene features into visual spatial description tasks by sampling topologically-diverse subgraphs from Go3D-S2G. |
| Outcome: | The proposed framework outperforms baselines on two VSD datasets and produces more spatially-diversified generation. |
OpenRLHF: A Ray-based Easy-to-use, Scalable and High-performance RLHF Framework (2025.emnlp-demos)
Copied to clipboard
Jian Hu, Xibin Wu, Wei Shen, Jason Klein Liu, Weixun Wang, Songlin Jiang, Haoran Wang, Hao Chen, Bin Chen, Wenkai Fang, null Xianyu, Yu Cao, Haotian Xu, Yiming Liu
| Challenge: | Existing RLHF frameworks face inference bottlenecks and complexity barriers restricting their accessibility for newcomers. |
| Approach: | They propose an open-source RLHF framework that can be used to train large language models. |
| Outcome: | The proposed framework achieves superior training efficiency with speedups ranging from 1.22 to 1.68 across different model sizes compared to state-of-the-art frameworks, while requiring significantly fewer lines of code for implementation. |
Activation-Guided Local Editing for Jailbreaking Attacks (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for jailbreaking Large Language Models (LLMs) are limited and produce incoherent or unreadable inputs. |
| Approach: | They propose a two-stage framework that performs a one-shot, scenario-based generation of context and rephrases the original malicious query to obscure its harmful intent. |
| Outcome: | The proposed framework achieves state-of-the-art Attack Success Rate, with gains of up to 37.74% over the strongest baseline, and excellent transferability to black-box and large-scale models. |
ERNIE-Layout: Layout Knowledge Enhanced Pre-training for Visually-rich Document Understanding (2022.findings-emnlp)
Copied to clipboard
Qiming Peng, Yinxu Pan, Wenjin Wang, Bin Luo, Zhenyu Zhang, Zhengjie Huang, Yuhui Cao, Weichong Yin, Yongfeng Chen, Yin Zhang, Shikun Feng, Yu Sun, Hao Tian, Hua Wu, Haifeng Wang
| Challenge: | Existing methods for visually rich document understanding lack layout-centered knowledge . experimental results show that ERNIE-Layout improves layout awareness . |
| Approach: | They propose a document pre-training solution with layout knowledge enhancement in the whole workflow to learn better representations that combine the features from text, layout, and image. |
| Outcome: | The proposed model outperforms existing models on key downstream tasks. |
Hyperlink-induced Pre-training for Passage Retrieval in Open-domain Question Answering (2022.acl-long)
Copied to clipboard
Jiawei Zhou, Xiaoguang Li, Lifeng Shang, Lan Luo, Ke Zhan, Enrui Hu, Xinyu Zhang, Hao Jiang, Zhao Cao, Fan Yu, Xin Jiang, Qun Liu, Lei Chen
| Challenge: | Existing methods to train dense passage retrieval have a large data gap between upstream and downstream relevance. |
| Approach: | They propose a method to pre-train the dense retriever with the text relevance induced by hyperlinks within Web documents. |
| Outcome: | The proposed method outperforms existing methods under different scenarios and in the open-domain question answering domain. |
How Chain-of-Thought Works? Tracing Information Flow from Decoding, Projection, and Activation (2026.findings-acl)
Copied to clipboard
| Challenge: | Chain-of-Thought (CoT) prompting significantly enhances model reasoning, yet its internal mechanisms remain poorly understood. |
| Approach: | They reversely traced information flow across decoding, projection, and activation phases and found that CoT may serve as a decoding space pruner . |
| Outcome: | The proposed framework can be used to design more efficient and robust prompts. |
NSP-BERT: A Prompt-based Few-Shot Learner through an Original Pre-training Task —— Next Sentence Prediction (2022.coling-1)
Copied to clipboard
| Challenge: | Recent studies have shown that using prompts to utilize language models to perform downstream tasks is more effective than using token-level methods such as PET. |
| Approach: | They propose to use a BERT original pre-training task abandoned by RoBERTa and other models to construct a sentence-level prompt-based method that does not need to fix the length of the prompt or the position to be predicted. |
| Outcome: | The proposed method performs better than PET and EFL on a BERT pre-training task and is comparable to other prompt-based methods. |
MIRTH: Mutual-Information Reasoning with Temporal Hubs for Vision-Language-Action Agents (2026.acl-long)
Copied to clipboard
| Challenge: | a recent study shows that VLA models suffer from temporal myopia that discards historical dynamics and reasoning gaps between high-level instructions and low-level motor commands. |
| Approach: | They propose a framework to address temporal myopia and autoregressive scalar decoding in VLAs . they propose two memory hubs that compress long-term scene evolution and short-term motion trends . |
| Outcome: | The proposed framework achieves state-of-the-art performance and exhibiting emergent error recovery capabilities. |