Papers by Pei Li

Tracing the Roots: A Multi-Agent Framework for Uncovering Data Lineage in Post-Training LLMs (2026.acl-long)

Challenge:	Large Language Models are increasingly utilized as role-playing agents to simulate personas in interactive settings.
Approach:	They propose a role-playing agent trained to explicitly ground responses in individual identity.
Outcome:	The proposed agent can generate persona-consistent responses in long-context dialogues while maintaining general instruction-following capabilities.

Copied to clipboard

Yu Li, Xiaoran Shang, Qizhi Pei, Yun Zhu, Xin Gao, Honglin Lin, Zhanping Zhong, Zhuoshi Pan, Zheng Liu, Xiaoyang Wang, Conghui He, Dahua Lin, Feng Zhao, Lijun Wu

Structured Data in NLP

Challenge:	High-quality post-training data is the primary engine driving LLM capabilities . datasets are often treated as isolated artifacts, overlooking their true developmental context .
Approach:	They propose a framework to reconstruct the evolutionary graph of dataset development using data lineage.
Outcome:	The proposed framework characterizes domain-specific structural patterns in Math-oriented datasets and general-domain corpora.

Mitigating Geospatial Knowledge Hallucination in Large Language Models: Benchmarking and Dynamic Factuality Aligning (2025.findings-emnlp)

Copied to clipboard

Shengyuan Wang, Jie Feng, Tianhui Liu, Dan Pei, Yong Li

An Empirical Study of Iterative Refinements for Non-autoregressive Translation (2025.acl-long)

Challenge:	Large language models (LLMs) have extensive world knowledge, but often generate inaccurate geospatial knowledge.
Approach:	They propose a framework for evaluation of large language models to mitigate hallucinations . they use Kahneman-Tversky Optimization to align LLMs with their reality .
Outcome:	The proposed evaluation framework uncovers hallucinations in 20 advanced LLMs.

Copied to clipboard

Yisheng Xiao, Pei Guo, Zechen Sun, Juntao Li, Kai Song, Min Zhang

ChartVerse: Scaling Chart Reasoning via Reliable Programmatic Synthesis from Scratch (2026.acl-long)

Challenge:	Iterative non-autoregressive (NAR) models have recently demonstrated impressive performance in varied generation tasks, surpassing the autoregressive Transformer.
Approach:	They propose a strategy to conduct efficient refinements without performance declines by using two simple metrics to identify potential problems existing in current refinement processes.
Outcome:	The proposed model outperforms the autoregressive Transformer by around one BLEU on average.

Copied to clipboard

Zheng Liu, Honglin Lin, Xiaoyang Wang, Xin Gao, Yu Li, Mengzhang Cai, Yun Zhu, Zhanping Zhong, Qizhi Pei, Zhuoshi Pan, Xiaoran Shang, Conghui He, Bin Cui, Wentao Zhang, Lijun Wu

CipherBank: Exploring the Boundary of LLM Reasoning Capabilities through Cryptography Challenge (2025.findings-acl)

Challenge:	Existing open-source vision language models lack high-quality training data for chart reasoning . current models are simplistic and repetitive, while associated QA pairs are prone to hallucinations .
Approach:	They propose a framework to synthesize complex charts and reliable reasoning data from scratch.
Outcome:	Experimental results show that ChartVerse-8B surpasses existing models in QA and difficulty . lack of high-quality training data hampers development of open-source models .

Copied to clipboard

Yu Li, Qizhi Pei, Mengyuan Sun, Honglin Lin, Chenlin Ming, Xin Gao, Jiang Wu, Conghui He, Lijun Wu

Aligning Large Language Models with Implicit Preferences from User-Generated Content (2025.acl-long)

Challenge:	Large language models (LLMs) have demonstrated remarkable capabilities, but their capabilities in cryptographic decryption tasks remain underexplored.
Approach:	They propose a benchmark to evaluate the reasoning capabilities of large language models in cryptographic decryption tasks.
Outcome:	The proposed benchmark examines the reasoning capabilities of large language models in cryptographic decryption tasks.

Copied to clipboard

Zhaoxuan Tan, Zheng Li, Tianyi Liu, Haodong Wang, Hyokun Yun, Ming Zeng, Pei Chen, Zhihan Zhang, Yifan Gao, Ruijie Wang, Priyanka Nigam, Bing Yin, Meng Jiang

Enhancing Translation Ability of Large Language Models by Leveraging Task-Related Layers (2024.lrec-main)

Challenge:	Existing preference learning methods rely heavily on curated data from humans or advanced LLMs, which is costly and difficult to scale.
Approach:	They propose a framework that leverages implicit preferences in unlabeled user-generated content to generate preference data.
Outcome:	The proposed framework transforms user-generated content into user queries and generates responses from the policy model.

Copied to clipboard

Pei Cheng, Xiayang Shi, Yinlin Li

Mitigating Lost in Multi-turn Conversation via Curriculum RL with Verifiable Accuracy and Abstention Rewards (2026.acl-long)

Challenge:	Experimental validation shows that adjusting task-related layers significantly improves performance on translation tasks while maintaining stability and accuracy on other tasks.
Approach:	They propose to adjust task-related layers in large models to better harness their machine translation capabilities by revealing the structure and characteristics of attention weights through singular value decomposition.
Outcome:	The proposed method reduces computational resource consumption and catastrophic forgetting while maintaining stability and accuracy on other tasks.

Copied to clipboard

Ming Li, Pei Chen, Zhenhao Zhang, Tao Yang, Xinyang Zhang, Han Li, Tianyu Cao, Ming Zeng, Zhuofeng Wu, Meng Jiang, Huasheng Li, Lihong Li, Bing Yin

Syllogistic Reasoning for Legal Judgment Analysis (2023.emnlp-main)

Challenge:	Large Language Models exhibit strong capabilities in single-turn instruction following but suffer from Lost-in-Conversation (LiC) when instructions are revealed progressively in multi-turn settings, models get "Lost in Conversation"
Approach:	They propose a framework that encourages models to generate correct answers and judge solvability in multi-turn conversations.
Outcome:	The proposed framework improves models' ability to balance problem-solving with abstention . it reduces premature answering behaviors that cause lost-in-conversation (LiC)

Copied to clipboard

Wentao Deng, Jiahuan Pei, Keyi Kong, Zhe Chen, Furu Wei, Yujun Li, Zhaochun Ren, Zhumin Chen, Pengjie Ren

ALERT: An LLM-powered Benchmark for Automatic Evaluation of Recommendation Explanations (2025.naacl-long)

Challenge:	Legal judgment assistants are developing fast due to impressive progress of large language models.
Approach:	They construct and manually correct a syllogistic reasoning dataset for legal judgment analysis using large language models as benchmarks.
Outcome:	The proposed dataset contains 11,239 criminal cases covering 4 criminal elements, 80 charges and 124 articles.

Copied to clipboard

Yichuan Li, Xinyang Zhang, Chenwei Zhang, Mao Li, Tianyi Liu, Pei Chen, Yifan Gao, Kyumin Lee, Kaize Ding, Zhengyang Wang, Zhihan Zhang, Jingbo Shang, Xian Li, Trishul Chilimbi

LEMMA: Learning from Errors for MatheMatical Advancement in LLMs (2025.findings-acl)

Challenge:	Existing benchmarks for recommendation explanation evaluation lack item diversity and user preferences data.
Approach:	They propose a model-agnostic recommendation explanation evaluation benchmark based on Amazon e-commerce categories with implicit preferences . they propose two novel automatic evaluators that enable scalable and human-preference aligned evaluation of explanations .
Outcome:	The proposed model-agnostic evaluation benchmark outperforms existing methods in a variety of domains.

Copied to clipboard

Zhuoshi Pan, Yu Li, Honglin Lin, Qizhi Pei, Zinan Tang, Wei Wu, Chenlin Ming, H. Vicky Zhao, Conghui He, Lijun Wu

ShieldLM: Empowering LLMs as Aligned, Customizable and Explainable Safety Detectors (2024.findings-emnlp)

Challenge:	Existing approaches focus on improving the quality of correct training data, neglecting the value contained in error data, thereby hindering the model’s reflective ability.
Approach:	They propose to improve LLM's reasoning ability by learning from error data and a grounded mistake augmentation method to collect representative errors.
Outcome:	The proposed model achieves significant performance improvements over other strong models with less than 90k data.

Copied to clipboard

Zhexin Zhang, Yida Lu, Jingyuan Ma, Di Zhang, Rui Li, Pei Ke, Hao Sun, Lei Sha, Zhifang Sui, Hongning Wang, Minlie Huang

Hephaestus: Improving Fundamental Agent Capabilities of Large Language Models through Continual Pre-Training (2025.naacl-long)

Challenge:	Existing tools for detecting safety issues in LLMs are expensive and inefficient.
Approach:	They propose an LLM-based safety detector which annotates the safety of queries and provides explanations for its decisions.
Outcome:	The proposed detector outperforms baselines on four sets of query-response pairs and is effective as a safety evaluator for advanced LLMs.

Copied to clipboard

Yuchen Zhuang, Jingfeng Yang, Haoming Jiang, Xin Liu, Kewei Cheng, Sanket Lokegaonkar, Yifan Gao, Qing Ping, Tianyi Liu, Binxuan Huang, Zheng Li, Zhengyang Wang, Pei Chen, Ruijie Wang, Rongzhi Zhang, Nasser Zalmout, Priyanka Nigam, Bing Yin, Chao Zhang

MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing (2026.acl-industry)

Challenge:	Existing LLMs often rely on complex prompting or extensive fine-tuning to introduce new capabilities while preserving strong generalizability.
Approach:	They propose a large-scale pre-training corpus to enhance LLM agents' capabilities . they use 103B agent-specific data encompassing 76,537 APIs .
Outcome:	The proposed training corpus outperforms open-source LLMs and commercial LLM agents on three agent benchmarks.

Copied to clipboard

Junbo Niu, Zheng Liu, Zhuangcheng Gu, Bin Wang, Linke Ouyang, Zhiyuan Zhao, Tao Chu, Tianyao He, Fan Wu, Qintong Zhang, Zhenjiang Jin, Guang Liang, Rui Zhang, Wenzheng Zhang, Yuan Qu, Zhifei Ren, Yuefeng Sun, Zirui Tang, Boyu Niu, Yuanhong Zheng, Dongsheng Ma, Ziyang Miao, Hejun Dong, Siyi Qian, Junyuan Zhang, Fangdong Wang, Jingzhou Chen, Xiaomeng Zhao, Liqun Wei, Wei Li, Shasha Wang, RuiLiang Xu, Yuanyuan Cao, Lu Chen, Qianqian Wu, Huaiyu Gu, Lindong Lu, Dechen Lin, null Shenguanlin, Xuanhe Zhou, Linfeng Zhang, Yuhang Zang, Xiaoyi Dong, Jiaqi Wang, Bo Zhang, Lei Bai, Pei Chu, Weijia Li, Jiang Wu, Lijun Wu, Zhenxiang Li, Guangyu Wang, Zhongying Tu, Chao Xu, Kai Chen, Bowen Zhou, Dahua Lin, Wentao Zhang, Conghui He

Green & Sustainable NLP

Challenge:	Document images are characterized by higher resolutions, denser content, and more complex structural layouts.
Approach:	They propose a 1.2B-parameter document parsing vision-language model that decouples layout analysis from local content recognition.
Outcome:	The proposed model surpasses general-purpose and domain-specific models on multiple benchmarks while maintaining significantly lower computational overhead.

ChatEdit: Towards Multi-turn Interactive Facial Image Editing via Dialogue (2023.emnlp-main)

Copied to clipboard

Xing Cui, Zekun Li, Pei Li, Yibo Hu, Hailin Shi, Chunshui Cao, Zhaofeng He

Visual Data in NLP

Challenge:	Existing approaches to interactive facial image editing treat multi-turn editing as a sequence of successive single-turn edits, leading to attribute forgetting and error accumulation.
Approach:	They propose a framework for interactive facial image editing through dialogues based on the CelebA-HQ dataset and a benchmark dataset to evaluate this.
Outcome:	The proposed framework outperforms existing methods and improves existing ones.

S2SPMN: A Simple and Effective Framework for Response Generation with Relevant Information (D18-1)

Copied to clipboard

Jiaxin Pei, Chenliang Li

Exploring Reversal Mathematical Reasoning Ability for Large Language Models (2024.findings-acl)

Challenge:	Existing work on how to generate relevant and informative responses is focusing on how dialogue systems generate information from large dialogue corpus.
Approach:	They propose to use dialogue corpus to generate relevant responses by using prototypes to extract semantic information from PMN.
Outcome:	The proposed model outperforms classical and strong baseline models in generating relevant and informative responses.

Copied to clipboard

Pei Guo, WangJie You, Juntao Li, Yan Bowen, Min Zhang

Instant Personalized Large Language Model Adaptation via Hypernetwork (2026.acl-long)

Challenge:	Large language models (LLMs) have been a success in the wide range of natural language understanding and reasoning tasks.
Approach:	They propose a training method to improve general and reversal reasoning abilities by using a reversed dataset.
Outcome:	The proposed method improves general and reversal reasoning abilities and alleviates the reverse curse.

Copied to clipboard

Zhaoxuan Tan, Zixuan Zhang, Haoyang Wen, Zheng Li, Rongzhi Zhang, Pei Chen, Fengran Mo, Zheyuan Liu, Qingkai Zeng, Qingyu Yin, Meng Jiang

IEvoAgent: Evolving Conversational Agent based on User Implicit Feedback (2026.acl-long)

Challenge:	Existing parameter-efficient fine-tuning methods require training a separate adapter for each user, making them computationally expensive and impractical for real-time updates.
Approach:	They propose a scalable framework that maps a user's profile directly to a full set of adapter parameters.
Outcome:	The proposed framework outperforms prompt-based personalization and OPPU while using substantially fewer computational resources at deployment.

Copied to clipboard

Yichen Cai, Jiayang Li, Junyuan Qiu, Jingya Guo, Weitao You, Changyuan Yang, Lingyun Sun, Pei Chen

UniConv: Unifying Retrieval and Response Generation for Large Language Models in Conversations (2025.acl-long)

Challenge:	Existing approaches to optimize conversational agents often rely on explicit preference pairs and expert evaluations.
Approach:	They propose a conversational agent framework that leverages the structured dependency between agent responses and user reactions to extract implicit feedback.
Outcome:	The proposed framework improves on MT-Bench-101, WildBench, and FB-Bech, and shows that mining implicit feedback supports better multi-turn alignment under evolving user preferences.

Copied to clipboard

Fengran Mo, Yifan Gao, Chuan Meng, Xin Liu, Zhuofeng Wu, Kelong Mao, Zhengyang Wang, Pei Chen, Zheng Li, Xian Li, Bing Yin, Meng Jiang

CTRLEval: An Unsupervised Reference-Free Metric for Evaluating Controlled Text Generation (2022.acl-long)

Challenge:	Existing conversational search systems are usually built with two different models . this separation restricts the system from leveraging the model's intrinsic knowledge simultaneously . Existing studies for developing unified models cannot fully address the aspects of understanding conversational context, managing retrieval independently, and generating responses.
Approach:	They propose to unify dense retrieval and response generation for large language models in conversation by fine-tuning and mitigating data discrepancy.
Outcome:	The proposed model can outperform existing models on five conversational search datasets and reduce inconsistency risks while mitigating data discrepancy.

Copied to clipboard

Pei Ke, Hao Zhou, Yankai Lin, Peng Li, Jie Zhou, Xiaoyan Zhu, Minlie Huang

Low-Resource NLP

Challenge:	Existing reference-free metrics have obvious limitations for evaluating controlled text generation models.
Approach:	They propose an unsupervised reference-free metric which evaluates controlled text generation from different aspects by formulating each aspect into multiple text infilling tasks.
Outcome:	The proposed metric has higher correlations with human judgments while obtaining better generalization of evaluating generated texts from different models and with different qualities.

Are ChatGPT and GPT-4 General-Purpose Solvers for Financial Text Analytics? A Study on Several Typical Tasks (2023.emnlp-industry)

Copied to clipboard

Xianzhi Li, Samuel Chan, Xiaodan Zhu, Yulong Pei, Zhiqiang Ma, Xiaomo Liu, Sameena Shah

ShopSimulator: Evaluating and Exploring RL-Driven LLM Agent for Shopping Assistants (2026.acl-long)

Challenge:	Recent large language models such as ChatGPT and GPT-4 have shown exceptional capabilities of generalist models . however, their applicability and effectiveness in specific domains like finance needs a better understanding .
Approach:	They conduct empirical studies to compare the performance of ChatGPT and GPT-4 on financial text analytical problems using eight benchmark datasets from five categories of tasks.
Outcome:	The proposed models outperform the state-of-the-art models on a wide range of financial text analytical tasks.

Copied to clipboard

Pei Wang, Yanan Wu, Xiaoshuai Song, Weixun Wang, Gengru Chen, Zhongwen Li, Kezhong Yan, Qi Liu, Ken Deng, Shuaibing Zhao, Shaopan Xiong, Xuepeng Liu, Xuefeng Chen, Wanxi Deng, Wenbo Su, Bo Zheng

Challenge:	Existing studies on large language model-based agents focus on evaluation benchmarks without training support.
Approach:	They propose a large-scale Chinese shopping simulation environment that uses large language models to train agents.
Outcome:	The proposed model performs poorly in a large-scale and challenging shopping environment in China.

Efficient Domain Adaptation for Non-Autoregressive Machine Translation (2024.findings-acl)

Copied to clipboard

WangJie You, Pei Guo, Juntao Li, Kehai Chen, Min Zhang

STORM-BORN: A Challenging Mathematical Derivations Dataset Curated via a Human-in-the-Loop Multi-Agent Framework (2025.findings-acl)

Challenge:	Existing non-parametric approaches like nearest neighbor machine translation have made small Autoregressive translation models less efficient . despite their impressive generalization and task performance, LLMs suffer from prohibitive inference cost when confronted with specific domains.
Approach:	They propose a domain adaptation approach that tailors a k-nearest-neighbor algorithm for NAT models that incorporates the parallel nature of NAT.
Outcome:	The proposed approach achieves significant improvements over the Base-NAT model and exhibits enhanced efficiency.

Copied to clipboard

Wenhao Liu, Zhenyi Lu, Xinyu Hu, Jerry Zhang, Dailin Li, Jiacheng Cen, Huilin Cao, Haiteng Wang, Yuhan Li, Xie Kun, Dandan Li, Pei Zhang, Chengbo Zhang, Yuxiang Ren, Xiaohong Huang, Yan Ma

LongLeader: A Comprehensive Leaderboard for Large Language Models in Long-context Scenarios (2025.naacl-long)

Challenge:	Existing datasets suffer from outdated and insufficient challenging content, neglecting human-like reasoning, and limited reliability due to single-LLM generation.
Approach:	They propose a human-in-the-loop, multi-agent data generation framework that integrates reasoning-dense filters, multiagent collaboration, and human mathematicians’ evaluations to ensure the reliability and quality of the dataset.
Outcome:	The proposed framework improves accuracy and quality of the 2,000-synthesized datasets by integrating reasoning-dense filters, multi-agent collaboration, and human mathematicians’ evaluations.

Copied to clipboard

Pei Chen, Hongye Jin, Cheng-Che Lee, Rulin Shao, Jingfeng Yang, Mingyu Zhao, Zhaoyu Zhang, Qin Lu, Kaiwen Men, Ning Xie, Huasheng Li, Bing Yin, Han Li, Lingyun Wang

ThinkAnswer Loss: Balancing Semantic Similarity and Exact Matching for LLM Reasoning Enhancement (2025.findings-emnlp)

Challenge:	LongLeader aims to assess different LLMs' long-context comprehension abilities . long-constext comprehension is a key bottleneck for many use cases .
Approach:	They propose a leaderboard to assess different LLMs' long-context comprehension abilities . they offer open-source access to the benchmarks and maintain a dedicated website .
Outcome:	The proposed model assesses different LLMs on selected benchmarks and provides open-source access to the benchmarks.

Copied to clipboard

Shan Yang, Kun Wu, Zeju Li, Linlin Zhang, Xiangyu Pei, Leike An, Yu Liu

Beyond Task-Level Context: Class-Conditional Context Vectors for Implicit In-Context Learning (2026.findings-acl)

Challenge:	Existing methods for knowledge distillation use Chain-of-Thought (CoT) and answer pairs, but they lack appropriate supervision signals.
Approach:	They propose a framework that decouples CoT and answer supervision . the framework applies semantic similarity constraints while maintaining strict literal matching for the answer .
Outcome:	The proposed framework decouples CoT and answer supervision while maintaining strict literal matching for the answer.

Copied to clipboard

Jianxin Zhang, Yilu Hu, Teng Liu, Pei Guo, Juntao Li

Trajectory2Task: Training Robust Tool-Calling Agents with Synthesized Yet Verifiable Data for Complex User Intents (2026.acl-long)

Challenge:	Existing approaches aggregate demonstrations from all classes into a shared, task-level context vector, capturing global task information but without explicitly preserving fine-grained, class-conditional semantic distinctions.
Approach:	They propose a class-conditional context vector extension to implicit in-context learning that explicitly models class-specific contextual information by constructing separate context vectors for each class.
Outcome:	The proposed extension outperforms task-level context vector baselines and achieves higher average accuracy than conventional few-shot learning on most models.

Copied to clipboard

Ziyi Wang, Yuxuan Lu, Yimeng Zhang, Pei Chen, Ziwei Dong, Jing Huang, Jiri Gesi, Xianfeng Tang, Chen Luo, Qun Liu, Yisi Sang, Hanqing Lu, Manling Li, Jin Lai, Dakuo Wang

Improving Large Language Models Function Calling and Interpretability via Guided-Structured Templates (2025.emnlp-main)

Challenge:	Tool-calling agents are increasingly deployed in real-world customer-facing workflows . but most studies on tool-callers focus on idealized settings with general, fixed, and well-specified tasks.
Approach:	They propose a tool-calling agent-based data pipeline that converts trajectories into user-facing tasks with controlled intent adaptations.
Outcome:	The proposed pipeline can be used to study tool use under three scenarios.

Copied to clipboard

Hy Dang, Tianyi Liu, Zhuofeng Wu, Jingfeng Yang, Haoming Jiang, Tao Yang, Pei Chen, Zhengyang Wang, Helen Wang, Huasheng Li, Bing Yin, Meng Jiang

Training Language Model to Critique for Better Refinement (2025.findings-acl)

Challenge:	Large language models (LLMs) have strong reasoning and tool-use capabilities, yet fail in real-world tool-interactions due to incorrect parameterization, poor tool selection, or misinterpretation of user intent.
Approach:	They propose a curriculum-inspired framework that leverages structured reasoning templates to guide LLMs through more deliberate step-by-step instructions for generating function calls.
Outcome:	The proposed framework reduces tool-use errors and improves interpretability and transparency of tool-using agents.

Copied to clipboard

Tianshu Yu, Chao Xiang, Mingchuan Yang, Pei Ke, Bosi Wen, Cunxiang Wang, Jiale Cheng, Li Zhang, Xinyu Mu, Chuxiong Sun, Minlie Huang

SynthFix: Adaptive Neuro-Symbolic Code Vulnerability Repair (2026.findings-acl)

Challenge:	Large language models (LLMs) have remarkable evaluation and critique capabilities, providing insightful feedback and identifying flaws in various tasks.
Approach:	They propose a framework to train critic models using refinement signals to generate feedback loops where critiques guide the model in refining its responses.
Outcome:	The proposed framework outperforms traditional methods and open-source models in terms of critique quality and refinement outcomes.

Copied to clipboard

Yifan Zhang, Jieyu Li, Kexin Pei, Yu Huang, Kevin Leach

Mosaic-IT: Cost-Free Compositional Data Synthesis for Instruction Tuning (2025.findings-acl)

Challenge:	Large Language Models (LLMs) struggle with complex semantic and structural correctness required for automated code repair.
Approach:	They propose a hybrid neural-symbolic framework that unifies code synthesis with compiler-informed symbolic feedback to improve LLM-based vulnerability repair.
Outcome:	The proposed framework improves code repair accuracy and efficiency over strong SFT and RFT training strategies on the FixJS and CodeFlaws benchmarks.

Copied to clipboard

Ming Li, Pei Chen, Chenguang Wang, Hongyu Zhao, Yijun Liang, YuPeng Hou, Fuxiao Liu, Tianyi Zhou

Rethinking the Alignment of Psychotherapy Dialogue Generation with Motivational Interviewing Strategies (2025.coling-main)

Challenge:	Current instruction tuning relies on teacher models or human intervention to generate and refine the instructions and responses for training, which are costly, non-sustainable, and may lack diversity.
Approach:	They propose a human/model-free compositional data synthesis method that can create rich and diverse augmentations from existing instruction tuning data to enhance large language models.
Outcome:	The proposed method improves performance over benchmarks and reduces training costs by 80% compared with original instruction tuning.

Copied to clipboard

Xin Sun, Xiao Tang, Abdallah El Ali, Zhuying Li, Pengjie Ren, Jan de Wit, Jiahuan Pei, Jos A.Bosch

PAD: A Robustness Enhancement Ensemble Method via Promoting Attention Diversity (2024.lrec-main)

Challenge:	Motivational interviewing (MI) is a client-centered counseling technique that encourages individuals to change behaviors through emphatic conversations.
Approach:	They propose to use large language models to generate more controllable dialogues with explainability by prompting LLMs to predict appropriate strategies as reasoning and utilizing these strategies to guide dialogue generation.
Outcome:	The proposed model generates more controllable and explainable dialogues with a set of MI skills.

Copied to clipboard

Yuting Yang, Pei Huang, Feifei Ma, Juan Cao, Jintao Li

Robustness in NLP

Challenge:	Existing approaches to enhance robustness of deep neural networks focus on perturbation . weak robustness is a problem for many types of adversarial attacks, authors say .
Approach:	They propose a lightweight framework for enhancing robustness by perturbing parameters of a model and diversifying adversarial example distributions among different models.
Outcome:	The proposed method can improve robustness against adversarial attacks while maintaining accuracy on clean data.

A Strategic Coordination Framework of Small LMs Matches Large LMs in Data Synthesis (2025.acl-long)

Copied to clipboard

Xin Gao, Qizhi Pei, Zinan Tang, Yu Li, Honglin Lin, Jiang Wu, Lijun Wu, Conghui He

Multi-modal Preference Alignment Remedies Degradation of Visual Instruction Tuning on Language Models (2024.acl-long)

Challenge:	Large Language Models suffer from high computational costs and environmental inefficiency . smaller LMs are more accessible and sustainable, but their individual capabilities often fall short . a collaborative framework for small LM combines specialized roles to iterative refinement and quality control .
Approach:	They propose a framework that aggregates specialized roles across small LMs to iterative refinement and quality control typically achieved by a single large LM.
Outcome:	The proposed framework aggregates specialized roles across small LMs to iterative refinement and quality control typically achieved by large LM.

Copied to clipboard

Shengzhi Li, Rongyu Lin, Shichao Pei

SAMem: State-Aware Memory as a Fine-Grained Memory for LLM Agents in Decision-Making (2026.findings-acl)

Challenge:	Multi-modal large language models (MLLMs) are expected to support multi-turn queries of interchanging image and text modalities in production.
Approach:	They propose to use visual-question-answering (VQA) datasets to annotate a 5k-sample VQA preference dataset and to investigate the degradation of VQA datasets.
Outcome:	The proposed model surpasses the instruction-following capabilities of the language model with DPO and SteerLM.

Copied to clipboard

Tong Wang, Pei Xu, Shiyue Cao, Likun Yang, Daipeng Li, Jianbin Jiao, Kaiqi Huang

Alleviating Over-smoothing for Unsupervised Sentence Representation (2023.acl-long)

Challenge:	Existing experiential memory approaches rely on task-level memory, but this lacks the situational alignment required for complex multi-step decision-making.
Approach:	They propose a new fine-grained memory paradigm that aligns memory retrieval with the current state instead of storing and reusing globally shared experiences.
Outcome:	Experiments on complex decision-making benchmarks show that the proposed state-aware memory outperforms existing experiential memory approaches and significantly improves task-solving efficiency.

Copied to clipboard

Nuo Chen, Linjun Shou, Jian Pei, Ming Gong, Bowen Cao, Jianhui Chang, Jia Li, Daxin Jiang

From Experts to Bases: Orthogonal Subspace Mixture for Continual Multimodal Instruction Tuning (2026.acl-long)

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.

Copied to clipboard

Pei Chen, Xilai Wang, null Shiqixu, Zejian Li, Lingyun Sun

MathFusion: Enhancing Mathematical Problem-solving of LLM through Instruction Fusion (2025.acl-long)

Challenge:	Existing parameter-efficient approaches to multimodal Continual Instruction Tuning suffer from knowledge interference and inefficient capacity expansion, limiting scalability.
Approach:	They propose a framework for multimodal Continual instruction tuning that decomposes adaptation weights into a globally shared pool of orthonormal bases to capture task-invariant knowledge.
Outcome:	Experiments show that MoBLoRA outperforms state-of-the-art methods while maintaining superior parameter efficiency.

Copied to clipboard

Qizhi Pei, Lijun Wu, Zhuoshi Pan, Yu Li, Honglin Lin, Chenlin Ming, Xin Gao, Conghui He, Rui Yan

MATINF: A Jointly Labeled Large-Scale Dataset for Classification, Question Answering and Summarization (2020.acl-main)

Challenge:	Large Language Models (LLMs) have shown impressive progress in mathematical problem-solving . current approaches to enhance mathematical reasoning focus on instance-level modifications .
Approach:	They propose a framework that enhances mathematical reasoning through cross-problem instruction synthesis.
Outcome:	The proposed framework boosts mathematical reasoning by 18.0 points while maintaining high data efficiency.

Copied to clipboard

Canwen Xu, Jiaxin Pei, Hongtao Wu, Yiyu Liu, Chenliang Li

Information Extraction & Text Mining

Challenge:	Existing datasets labeled for one task hinder multi-task learning . task-specific data make models learn task-related leakage features rather than meaningful knowledge that could generalize to other tasks.
Approach:	They propose to jointly label large-scale NLP dataset MATINF . it contains 1.07 million question-answer pairs with human-labeled categories .
Outcome:	The proposed dataset is applicable for classification, question answering, and summarization.

Structural Contrastive Pretraining for Cross-Lingual Comprehension (2023.findings-acl)

Copied to clipboard

Nuo Chen, Linjun Shou, Tengtao Song, Ming Gong, Jian Pei, Jianhui Chang, Daxin Jiang, Jia Li

DSRAG: A Double-Stream Retrieval-Augmented Generation Framework for Countless Intent Detection (2025.naacl-industry)

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.

Copied to clipboard

Pei Guo, Enjie Liu, Ruichao Zhong, Mochi Gao, Yunzhi Tan, Bo Hu, Zang Li

DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models (2025.naacl-long)

Challenge:	Current intent detection work experiments with minor intent categories.
Approach:	They propose a retrieval-augmented generation framework that uses query-to-query and query- to-metadata approaches to retrieve intents from metadata.
Outcome:	The proposed framework improves on query-to-query (Q2Q) and query- to-metadata (Q 2M) approaches.

Copied to clipboard

Jianyu Liu, Hangyu Guo, Ranjie Duan, Xingyuan Bu, Yancheng He, Shilong Li, Hui Huang, Jiaheng Liu, Yucheng Wang, Chenchen Jing, Xingwei Qu, Xiao Zhang, Pei Wang, Yanan Wu, Jihao Gu, Yangguang Li, Jianke Zhu

When to Use What: An In-Depth Comparative Empirical Analysis of OpenIE Systems for Downstream Applications (2023.acl-long)

Challenge:	Multimodal Large Language Models (MLLMs) pose unique safety challenges due to their integration of visual and textual data.
Approach:	They propose a method to disentangle risks through step-by-step reasoning within multimodal inputs.
Outcome:	The proposed approach improves safety alignment in MLLMs by fine-tuning and iterative Reinforcement Learning from AI feedback.

Copied to clipboard

Kevin Pei, Ishan Jindal, Kevin Chen-Chuan Chang, ChengXiang Zhai, Yunyao Li

Open Information Extraction

Challenge:	Existing consensus on which OpenIE model is best for each application is lacking . different assumptions made by different models and datasets have a statistically significant effect on performance, making it important to choose the most appropriate OpenIE system for one’s applications.
Approach:	They propose to use OpenIE to extract relation tuples from plain text to compare different models and training sets to find the best model for their applications.
Outcome:	The proposed models perform well on a Complex QA application.

Doc-V*: Coarse-to-Fine Interactive Visual Reasoning for Multi-Page Document VQA (2026.acl-long)

Copied to clipboard

Yuanlei Zheng, Pei Fu, Hang Li, Ziyang Wang, Yuyi Zhang, Wenyu Ruan, Xiaojin Zhang, Zhongyu Wei, Zhenbo Luo, Jian Luan, Wei Chen, Xiang Bai

Unveiling the Implicit Toxicity in Large Language Models (2023.emnlp-main)

Challenge:	Existing OCR-free approaches to document visual question answering are brittle and passive.
Approach:	They propose an OCR-free agentic framework that casts multi-page DocVQA as sequential evidence aggregation.
Outcome:	The proposed framework outperforms open-source and proprietary models in five benchmarks and improves out-of-domain performance by 47.9% over baseline.

Copied to clipboard

Jiaxin Wen, Pei Ke, Hao Sun, Zhexin Zhang, Chengfei Li, Jinfeng Bai, Minlie Huang

MetaLadder: Ascending Mathematical Solution Quality via Analogical-Problem Reasoning Transfer (2025.findings-emnlp)

Challenge:	Recent studies focus on probing toxic outputs that can be easily detected with existing toxicity classifiers, but LLMs can generate diverse implicit toxic output that are difficult to detect via simply zero-shot prompting.
Approach:	They propose a reinforcement learning based attacking method to induce the implicit toxic outputs in large language models by fine-tuning toxicity classifiers.
Outcome:	The proposed method generates implicit toxic outputs that are difficult to detect via zero-shot prompting on five widely-adopted toxicity classifiers.

Copied to clipboard

Honglin Lin, Zhuoshi Pan, Qizhi Pei, Xin Gao, Yu Li, Mengzhang Cai, Conghui He, Lijun Wu

EviReport: From Reasoned Outlines to Evidence Tracked Long-Form Reports (2026.findings-acl)

Challenge:	Current paradigms generate CoT and answers directly for a given problem, diverging from human problem-solving strategies to some extent.
Approach:	They propose a framework that explicitly prompts LLMs to recall and reflect on meta-problems alongside their CoT solutions before addressing the target problem.
Outcome:	The proposed framework outperforms standard CoT-based methods on mathematical benchmarks and significantly improves their reasoning accuracy.

Copied to clipboard

Zihan Liu, Jianhui li, Zexin Wang, Fei Sun, Jingjing LI, Zheyuan Li, Ke Xiang, Hang Cui, Houhua Gong, Changhua Pei, Gaogang Xie

Isotropy-Enhanced Conditional Masked Language Models (2023.findings-emnlp)

Challenge:	Evidence-intensive reports often produce fluent but under-supported drafts . eviReport is an evidence-grounded workflow for automated long-form report generation .
Approach:	They propose an evidence-tracked workflow that organizes corpus evidence into compact, traceable units and retrieves query-relevant subgraphs into retrieval-ready packages.
Outcome:	The proposed workflow outperforms baselines in factual coverage, factual accuracy and visual evidence integration.

Copied to clipboard

Pei Guo, Yisheng Xiao, Juntao Li, Yixin Ji, Min Zhang

Detecting Customer Complaint Escalation with Recurrent Neural Networks and Manually-Engineered Features (N19-2)

Challenge:	Existing non-autoregressive models with auto-regressing decoding paradigms have been used for various text generation tasks to accelerate inference but at the cost of generation quality to some extent.
Approach:	They propose to use Look Neighbors strategy to enhance learning of target token representations during training to achieve a good balance between inference speedup and generation quality.
Outcome:	The proposed models outperform current models on 4 WMT datasets and outperformed the current SoTA results.

Copied to clipboard

Wei Yang, Luchen Tan, Chunwei Lu, Anqi Cui, Han Li, Xi Chen, Kun Xiong, Muzi Wang, Ming Li, Jian Pei, Jimmy Lin

Challenge:	e-commerce companies often have the option of escalating complaints by filing grievances with a government authority . this is detrimental to an ecommerce company, but this problem is challenging to solve by integrating recurrent neural networks with manually-engineered features.
Approach:	They propose a model that integrates recurrent neural networks with manually-engineered features to identify cases where the customer expresses such an intent.
Outcome:	The proposed model outperforms baseline models and provides better recall and triage for specialized agents.

Pre-train and Plug-in: Flexible Conditional Text Generation with Variational Auto-Encoders (2020.acl-main)

Copied to clipboard

Yu Duan, Canwen Xu, Jiaxin Pei, Jialong Han, Chenliang Li

ROSE: An Intent-Centered Evaluation Metric for NL2SQL (2026.acl-long)

Challenge:	Existing conditional generation models cannot handle emerging conditions due to their joint end-to-end learning fashion.
Approach:	They propose a framework for conditional text generation that decouples the text generation module from the condition representation module to allow "one-to-many" conditional generation.
Outcome:	The proposed framework decouples the text generation module from the condition representation module to allow “one-to-many” conditional generation.

Copied to clipboard

Wenqi Pei, Shizheng Hou, Boyan Li, Chen Han, Zhichao Shi, Yuyu Luo

Programming Languages in NLP

Challenge:	Existing evaluation metrics for evaluating the effectiveness of Natural Language to SQL (NL2SQL) solutions are becoming unreliable due to its sensitiveness to syntactic variation and inconsistent consistency with ground-truth SQL.
Approach:	They propose an intent-centered metric that focuses on whether the predicted SQL answers the question, rather than consistency with the ground-truth SQL.
Outcome:	The proposed metric outperforms the next-best metric by nearly 24% on the expert-aligned validation set ROSE-VEC.

SWAG: Storytelling With Action Guidance (2024.findings-emnlp)

Copied to clipboard

Jonathan Pei, Zeeshan Patel, Karim El-Refai, Tianle Li

More than Minorities and Majorities: Understanding Multilateral Bias in Language Generation (2024.findings-acl)

Challenge:	Large language models (LLMs) are used for one-shot creation, but they can produce inconsistent but not necessarily engaging content.
Approach:	They propose a novel approach to storytelling with large language models that reduces story writing to a search problem through a two-model feedback loop.
Outcome:	The proposed approach outperforms existing methods when evaluated by GPT-4 and through human evaluation.

Copied to clipboard

Jiaxu Zhao, Zijing Shi, Yitong Li, Yulong Pei, Ling Chen, Meng Fang, Mykola Pechenizkiy

Transferable Direct Prompt Injection via Activation-Guided MCMC Sampling (2025.emnlp-main)

Challenge:	Existing studies on bias dataset construction and mitigation focus on one demographic group . in real-world applications, there are more than two demographic groups at risk of the same bias.
Approach:	They propose to analyze and reduce biases across multiple demographic groups using a multi-demographic bias dataset.
Outcome:	The proposed method can mitigate biases among multiple demographic groups effectively, the authors show .

Copied to clipboard

Minghui Li, Hao Zhang, Yechao Zhang, Wei Wan, Shengshan Hu, Pei Xiaobing, Jing Wang

Multi-Modal Sarcasm Detection via Cross-Modal Graph Convolutional Network (2022.acl-long)

Challenge:	Experimental results show superior cross-model transferability . Prompt injection attacks are among the most critical threats .
Approach:	They propose an activations-guided prompt injection attack framework to address the impracticality of existing white-box/gray-box methods and the poor transferability of black-box approaches.
Outcome:	The proposed framework achieves 49.6% success rate and 34.6% improvement over human-crafted prompts on five mainstream LLMs.

Copied to clipboard

Bin Liang, Chenwei Lou, Xiang Li, Min Yang, Lin Gui, Yulan He, Wenjie Pei, Ruifeng Xu