Papers by Junjie Wang

From Allies to Adversaries: Manipulating LLM Tool-Calling through Adversarial Injection (2025.naacl-long)

Challenge:	Current multimodal large language models (MLLMs) show limited understanding of dental images.
Approach:	They propose a dental-specialized multimodal large language model trained via staged multimodal alignment and reinforcement learning.
Outcome:	The proposed model outperforms state-of-the-art models on disease classification and dental VQA tasks.

Copied to clipboard

Rupeng Zhang, Haowei Wang, Junjie Wang, Mingyang Li, Yuekai Huang, Dandan Wang, Qing Wang

Repairing Catastrophic-Neglect in Text-to-Image Diffusion Models via Attention-Guided Feature Enhancement (2024.findings-emnlp)

Challenge:	Toolcalling has changed Large Language Model (LLM) applications by integrating external tools, but it also introduces new security vulnerabilities, particularly in the tool scheduling mechanisms of LLM, which have not been extensively studied.
Approach:	They propose a framework that exploits vulnerabilities in Large Language Models through adversarial tool injection to execute privacy theft, launch denial-of-service attacks, and manipulate business competition.
Outcome:	The proposed framework exploits vulnerabilities in LLM tool-calling systems through adversarial tool injection.

Copied to clipboard

Zhiyuan Chang, Mingyang Li, Junjie Wang, Yi Liu, Qing Wang, Yang Liu

AgentGym2: Benchmarking Large Language Model Agents in De-Idealized Real-World Environments (2026.acl-long)

Challenge:	Text-to-Image Diffusion models generate high-quality images from textual descriptions, but they often produce images that do not fully align with the input prompts, resulting in semantic inconsistencies.
Approach:	They propose an automated repair approach to address catastrophic-neglect in T2I DMs.
Outcome:	The proposed model achieves 10.1%-16.3% higher Correct Rate in image generation compared to baselines.

Copied to clipboard

Zhiheng Xi, Dingwen Yang, Jiaqi Liu, Jixuan Huang, Honglin Guo, Baodai Huang, Tinggang Chen, Qi Zhang, Zhonghang Lu, Chenyu Liu, Jiajun Sun, Jiazheng Zhang, Dingwei Zhu, Xin Guo, Junzhe Wang, Zhihao Zhang, Yuming Yang, Junjie Ye, Minghe Gao, Dongrui Liu, Jiaming Ji, Guohao Li, Tao Gui, Qi Zhang, Xuanjing Huang

CodeExp: Explanatory Code Document Generation (2022.findings-emnlp)

Challenge:	Existing benchmarks evaluate agents in simplified, idealized settings, relying on pre-packaged tool interfaces, overlooking critical steps, and assume inputs are clean and fully specified.
Approach:	They propose a framework that evaluates language agents in simplified, idealized settings . they show that even SOTA systems like Gemini and GPT-5 struggle on AgentGym2 .
Outcome:	Experiments on 15 proprietary and open-source models show that even SOTA systems like Gemini and GPT-5 struggle on AgentGym2 .

Copied to clipboard

Haotian Cui, Chenglong Wang, Junjie Huang, Jeevana Priya Inala, Todd Mytkowicz, Bo Wang, Jianfeng Gao, Nan Duan

Explainability & Interpretability in NLP

Challenge:	Existing code-to-text generation models produce only high-level code summaries that do not capture implementation-level choices essential for these scenarios.
Approach:	They propose a code explanation generation task that uses code docstrings to refine models.
Outcome:	The proposed model can generate well-structured long docstrings comparable to human-written ones.

Causal Intervention Improves Implicit Sentiment Analysis (2022.coling-1)

Copied to clipboard

Siyin Wang, Jie Zhou, Changzhi Sun, Junjie Ye, Tao Gui, Qi Zhang, Xuanjing Huang

Sentiment Analysis

Challenge:	Existing neural models struggle with implicit sentiment analysis because they latch onto spurious correlations, resulting in poor generalization and robustness.
Approach:	They propose a CausaL intervention model for implicit sEntiment ANalysis using instrumental variable to eliminate confounding causal effects and extract the pure causal effect between sentence and sentiment.
Outcome:	The proposed model extracts the pure causal effect between sentence and sentiment using instrumental variable.

WebClipper: Efficient Evolution of Web Agents with Graph-based Trajectory Pruning (2026.acl-long)

Copied to clipboard

Junjie Wang, Zequn Xie, Dan Yang, Jie Feng, Yue Shen, Duolin Sun, Meixiu Long, Yihan Jiao, Zhehao Tan, Jian Wang, Peng Wei, Jinjie Gu

Structured Data in NLP

Challenge:	Open-source web agents rely on long tool-call trajectories with cyclic reasoning loops and exploration of unproductive branches.
Approach:	They propose a framework that compresses web agent trajectories via graph-based pruning.
Outcome:	The proposed framework reduces tool-call rounds by 20% while improving accuracy and efficiency while maintaining the same level of performance as existing models.

Generative Text-to-Image Retrieval via Hierarchical Identifiers and Semantic Internalization (2026.findings-acl)

Copied to clipboard

Jie Huang, Junjie Wang, Xin Liao, Ziyou Jiang, Wenshuo Wang, Shoubin Li, Qing Wang

ToolBeHonest: A Multi-level Hallucination Diagnostic Benchmark for Tool-Augmented Large Language Models (2024.emnlp-main)

Challenge:	Existing text-to-image retrieval methods suffer from limited semantic discriminability, alignment bias, and closed-set restrictions.
Approach:	They propose a framework for semantic internalization for Generative Multimodal Alignment . they construct multi-granularity hierarchical identifiers to ensure unique, semantically consistent image representations .
Outcome:	The proposed framework outperforms state-of-the-art frameworks on Flickr30K and MS-COCO datasets . it achieves average Recall@1, Recall @5, and Recall_10 improvements of 10.65%, 8.50%, and 7.00% .

Copied to clipboard

Yuxiang Zhang, Jing Chen, Junjie Wang, Yaxin Liu, Cheng Yang, Chufan Shi, Xinyu Zhu, Zihao Lin, Hanwen Wan, Yujiu Yang, Tetsuya Sakai, Tian Feng, Hayato Yamana

Improving Discriminative Capability of Reward Models in RLHF Using Contrastive Learning (2024.emnlp-main)

Challenge:	Currently, tool-augmented large language models (LLMs) only achieve total scores of 45.3 and 37.0, respectively, on a scale of 100.
Approach:	They propose a multi-level diagnostic process to assess the LLM's hallucinations through two perspectives: depth and breadth.
Outcome:	The proposed diagnostic process assesses the hallucinations of large language models through two perspectives: depth and breadth.

Copied to clipboard

Lu Chen, Rui Zheng, Binghai Wang, Senjie Jin, Caishuang Huang, Junjie Ye, Zhihao Zhang, Yuhao Zhou, Zhiheng Xi, Tao Gui, Qi Zhang, Xuanjing Huang

Learning to Plan for Retrieval-Augmented Large Language Models from Knowledge Graphs (2024.findings-emnlp)

Challenge:	Current methods rely on ranking losses to teach reward model to assess preferences, but they are susceptible to noise and ambiguous data, often failing to deeply understand human intentions.
Approach:	They propose a method that incorporates contrastive learning into the reward modeling process to enhance generalization and stabilize the reinforcement learning training process.
Outcome:	The proposed method enhances generalization of the reward model, stabilizes the reinforcement learning training process, and improves the final alignment with human preferences.

Copied to clipboard

Junjie Wang, Mingyang Chen, Binbin Hu, Dan Yang, Ziqi Liu, Yue Shen, Peng Wei, Zhiqiang Zhang, Jinjie Gu, Jun Zhou, Jeff Pan, Wen Zhang, Huajun Chen

DARM: Distribution-Aware Reward Modeling by Alleviating Biases from Low Preference-Context Dependency Data (2026.acl-long)

Challenge:	Recent studies have attempted to enhance the performance of large language models (LLMs) in complex question-answering (QA) tasks by combining step-wise planning with external retrieval.
Approach:	They propose a framework for enhancing LLMs’ planning capabilities by using planning data derived from knowledge graphs (KGs).
Outcome:	The proposed framework improves LLMs’ planning capabilities by using knowledge graphs (KGs) the proposed framework is compared with existing frameworks on multiple datasets and shows that it is effective for large language models.

Copied to clipboard

Shaofan Liu, Guoqiang Zhang, Shihan Dou, Huiyuan Zheng, Yiming Zhou, Junjie Ye, Shaowen Wang, Shichun Liu, Jiazheng Zhang, Tao Gui, Qi Zhang, Xuanjing Huang

Challenge:	Existing methods for training reward models are vulnerable to context neglect and degraded accuracy.
Approach:	They propose distribution-aware reward modeling that augments the RM objective with a conditional mutual information regularizer that maximizes context and the predicted reward conditioned on the response.
Outcome:	The proposed model improves performance in RLHF and improves accuracy in other settings.

RATE-Nav: Region-Aware Termination Enhancement for Zero-shot Object Navigation with Vision-Language Models (2025.findings-acl)

Copied to clipboard

Junjie Li, Nan Zhang, Xiaoyang Qu, Kai Lu, Guokuan Li, Jiguang Wan, Jianzong Wang

DEFT: Demystifying VLN Failures via a Unified Dual-View Explainability Framework for LLM-based Agents (2026.acl-long)

Challenge:	Object navigation is a fundamental task in embodied artificial intelligence.
Approach:	They propose a region-aware Termination-Enhanced method that incorporates visual language models and exploration rates to enable efficient termination.
Outcome:	The proposed method achieves a success rate of 67.8% and an SPL of 31.3% on the HM3D dataset.

Copied to clipboard

Yawen Wang, Yihan Dai, Jianming Chen, Junjie Wang, Qing Wang

Explainability & Interpretability in NLP

Challenge:	Existing interpretability methods isolate temporal criticality from feature salience, creating an alignment gap and failing to account for the behavioral instability of black-box agents.
Approach:	They propose a unified dual-view framework that jointly analyzes when a decision is pivotal and what visual evidence grounds it.
Outcome:	Extensive experiments on MatterPort3D show that DEFT outperforms baselines in both temporal and feature fidelity.

AdapterDistillation: Non-Destructive Task Composition with Knowledge Distillation (2023.emnlp-industry)

Copied to clipboard

Junjie Wang, Yicheng Chen, Wangshu Zhang, Sen Hu, Teng Xu, Jing Zheng

Green & Sustainable NLP

Challenge:	Recent work on learning from multiple tasks has shown that adding an extra fusion layer to implement knowledge composition is non-scalable for some applications.
Approach:	They propose a two-stage knowledge distillation algorithm to extract task specific knowledge by using local data to train a student adapter.
Outcome:	Experiments on frequently asked question retrieval in task-oriented dialog systems validate the efficiency of AdapterDistillation.

STAD: Self-Training with Ambiguous Data for Low-Resource Relation Extraction (2022.coling-1)

Copied to clipboard

Junjie Yu, Xing Wang, Jiangjiang Zhao, Chunjie Yang, Wenliang Chen

Agentic Episodic Control (2026.findings-acl)

Challenge:	Existing approaches for low-resource relation extraction use only confident instances and uncertain instances.
Approach:	They propose a self-training approach for low-resource relation extraction using auto-annotated instances.
Outcome:	The proposed method improves on two widely used datasets with low-resource settings.

Copied to clipboard

Xidong Yang, Wenhao Li, Junjie Sheng, Yun Hua, Haosheng Chen, Chuyun Shen, Xiangfeng Wang

Green & Sustainable NLP

Challenge:	Existing methods for reinforcement learning (RL) are limited by poor data efficiency and weak generalization.
Approach:	They propose a novel architecture that integrates large language models into episodic RL.
Outcome:	The proposed architecture achieves 2–6 higher data efficiency than baselines and is the only method to solve complex tasks like UnlockLocal with over 90% success.

MegaPairs: Massive Data Synthesis for Universal Multimodal Retrieval (2025.acl-long)

Copied to clipboard

Junjie Zhou, Yongping Xiong, Zheng Liu, Ze Liu, Shitao Xiao, Yueze Wang, Bo Zhao, Chen Jason Zhang, Defu Lian

Zero-Shot Learners for Natural Language Understanding via a Unified Multiple Choice Perspective (2022.emnlp-main)

Challenge:	despite the growing demand for multimodal retrieval, there is a lack of training data.
Approach:	They propose a data synthesis method that leverages vision language models and open-domain images to generate high-quality data.
Outcome:	The proposed method outperforms baseline models on 70 more datasets and can scale up.

Copied to clipboard

Ping Yang, Junjie Wang, Ruyi Gan, Xinyu Zhu, Lin Zhang, Ziwei Wu, Xinyu Gao, Jiaxing Zhang, Tetsuya Sakai

Low-Resource NLP

Challenge:	Existing approaches to zero-shot learning are format-agnostic and can address new learning tasks without additional training.
Approach:	They propose a new paradigm for zero-shot learning that is format agnostic and compatible with any format and applicable to a list of language tasks.
Outcome:	The proposed model shows state-of-the-art performance on several benchmarks and produces satisfactory results on tasks such as text classification and commonsense reasoning.

MulDimIF: A Multi-Dimensional Constraint Framework for Evaluating and Improving Instruction Following in Large Language Models (2026.findings-acl)

Copied to clipboard

Junjie Ye, Caishuang Huang, Zhuohan Chen, Wenjie Fu, Chenyuan Yang, Leyi Yang, Yilong Wu, Peng Wang, Meng Zhou, Xiaolong Yang, Tao Gui, Qi Zhang, Zhongchao Shi, Jianping Fan, Xuanjing Huang

APIRecX: Cross-Library API Recommendation via Pre-Trained Language Model (2021.emnlp-main)

Challenge:	Existing research has focused on constraint categories, offering little guidance for improving instruction following abilities.
Approach:	They propose a multi-dimensional constraint framework that allows for instruction following . they construct 9,106 code-verifiable samples and evaluate 18 LLMs .
Outcome:	The proposed framework improves instruction following performance without compromising general performance.

Copied to clipboard

Yuning Kang, Zan Wang, Hongyu Zhang, Junjie Chen, Hanmo You

MIRTT: Learning Multimodal Interaction Representations from Trilinear Transformers for Visual Question Answering (2021.findings-emnlp)

Challenge:	API recommendation tools can help programmers use APIs by recommending which APIs to be used next given the APIs that have been written.
Approach:	They propose a cross-library API recommendation approach that uses BPE to split API calls in each sequence and pre-train a GPT based language model.
Outcome:	The proposed APIRecX can recommend APIs that are previously regarded as OOV . it can migrate knowledge of existing libraries to a new library and recommend API that is previously viewed as OVO .

Copied to clipboard

Junjie Wang, Yatai Ji, Jiaqi Sun, Yujiu Yang, Tetsuya Sakai

Beyond Boundaries: Learning a Universal Entity Taxonomy across Datasets and Languages for Open Named Entity Recognition (2025.coling-main)

Challenge:	Existing bilinear methods focus on inter-modality information between images and questions . existing models focus on the interaction between images, questions, and images .
Approach:	They propose a trilinear interaction framework that incorporates attention mechanisms for capturing inter-modality and intra-modal relationships.
Outcome:	The proposed model outperforms bilinear models on the Visual7W Telling task and VQA-1.0 Multiple Choice task and outperformed baselines on the VQA, TDIUC and GQA datasets.

Copied to clipboard

Yuming Yang, Wantong Zhao, Caishuang Huang, Junjie Ye, Xiao Wang, Huiyuan Zheng, Yang Nan, Yuran Wang, Xueying Xu, Kaixin Huang, Yunke Zhang, Tao Gui, Qi Zhang, Xuanjing Huang

Play Guessing Game with LLM: Indirect Jailbreak Attack with Implicit Clues (2024.findings-acl)

Challenge:	Current Large Language Models struggle with complex entity taxonomies in open domains and lack NER capabilities.
Approach:	They propose a dataset to guide LLMs' generalization in Open NER under a universal entity taxonomy.
Outcome:	The proposed model outperforms GPT-4 in 3 out-of-domain benchmarks across 15 datasets and 6 languages.

Copied to clipboard

Zhiyuan Chang, Mingyang Li, Yi Liu, Junjie Wang, Qing Wang, Yang Liu

Measuring Data Diversity for Instruction Tuning: A Systematic Analysis and A Reliable Metric (2025.acl-long)

Challenge:	Existing jailbreak attacks primarily utilize scenario camouflage techniques, however their explicit mention of malicious intent will be easily recognized and defended by LLMs.
Approach:	They propose an indirect jailbreak attack approach, Puzzler, which can bypass LLM’s defensive strategies and obtain malicious response by implicitly providing LLMs with some clues about the original malicious query.
Outcome:	The proposed approach can bypass the LLM’s defensive strategies and obtain malicious response by implicitly providing LLMs with some clues about the original malicious query.

Copied to clipboard

Yuming Yang, Yang Nan, Junjie Ye, Shihan Dou, Xiao Wang, Shuo Li, Huijie Lv, Tao Gui, Qi Zhang, Xuanjing Huang

PHMOSpell: Phonological and Morphological Knowledge Guided Chinese Spelling Check (2021.acl-long)

Challenge:	Existing studies have explored various diversity-aware data selection methods to construct high-quality datasets and enhance model performance.
Approach:	They propose to use data diversity to measure instruction tuning of large language models.
Outcome:	The proposed diversity metric outperforms existing methods on simulated and real-world data and shows that it captures diversity variations and achieves a 0.97 correlation with instruction tuning.

Copied to clipboard

Li Huang, Junjie Li, Weiwei Jiang, Zhiyu Zhang, Minchuan Chen, Shaojun Wang, Jing Xiao

Syntactic Text Processing

Challenge:	False gram and phonological errors make Chinese spelling check difficult . a novel end-to-end trainable model outperforms existing methods .
Approach:	They propose a trainable Chinese spelling check model that integrates phonological and visual information into a pre-trained language model.
Outcome:	The proposed model outperforms existing state-of-the-art models on three benchmarks.

Mimicking the Familiar: Dynamic Command Generation for Information Theft Attacks in LLM Tool-Learning System (2025.acl-long)

Copied to clipboard

Ziyou Jiang, Mingyang Li, Guowei Yang, Junjie Wang, Yuekai Huang, Zhiyuan Chang, Qing Wang

RethinkingTMSC: An Empirical Study for Target-Oriented Multimodal Sentiment Classification (2023.findings-emnlp)

Challenge:	Existing approaches to attack Large Language Model (LLM) tool-learning systems are black-box oriented and rely on static commands that cannot adapt flexibly to the changes in user queries and the invocation chain.
Approach:	They propose a dynamic attack comment generation approach for information theft attacks in LLM tool-learning systems that mimics the familiar by inferring the information utilized by upstream tools.
Outcome:	The proposed approach outperforms baselines with +13.2% ASRTheft and can be generalized to new tool-learning systems to expose their information leakage risks.

Copied to clipboard

Junjie Ye, Jie Zhou, Junfeng Tian, Rui Wang, Qi Zhang, Tao Gui, Xuanjing Huang

Information Retrieval

Challenge:	Recent studies have shown that current TMSC systems rely on textual information, and the progress in tackling this task has slowed down.
Approach:	They propose to integrate both visual and textual information to improve the performance of TMSC by considering multimodal information.
Outcome:	The proposed model integrates both visual and textual information to improve performance.

FocusLLM: Precise Understanding of Long Context by Dynamic Condensing (2025.acl-long)

Copied to clipboard

Zhenyu Li, Yike Zhang, Tengyu Pan, Yutao Sun, Zhichao Duan, Junjie Fang, Rong Han, Zixuan Wang, Jianyong Wang

Challenge:	Existing context condensing methods cannot accurately understand the full context, as there is a considerable amount of information loss in the condensed process.
Approach:	They propose a framework to extend the fixed context length of any decoder-only LLM by distilling crucial information from long sequences.
Outcome:	The proposed framework extends the fixed context length of any decoder-only LLM, allowing it to focus on relevant information from very long sequences.

Learning to Detect Noisy Labels Using Model-Based Features (2022.findings-emnlp)

Copied to clipboard

Zhihao Wang, Zongyu Lin, Junjie Wen, Xianxin Chen, Peiqi Liu, Guidong Zheng, Yujun Chen, Zhilin Yang

Domain Differential Adaptation for Neural Machine Translation (D19-56)

Challenge:	Existing approaches to reduce label noise rely on heuristics and sample losses.
Approach:	They propose a method that transfers the noise distribution to a clean set and trains a model to distinguish noisy labels from clean ones using model-based features.
Outcome:	Empirically, the proposed approach improves over strong baselines on a wide range of tasks including text classification and speech recognition.

Copied to clipboard

Zi-Yi Dou, Xinyi Wang, Junjie Hu, Graham Neubig

Machine Translation

Challenge:	Neural networks are data hungry and domain sensitive, so it is difficult to obtain labeled data for every domain.
Approach:	They propose a framework for domain adaptation where we model the difference between domains instead of smoothing over them.
Outcome:	The proposed framework improves on domain adaptation in multiple experimental settings.

GRAFF: GRaph-Augmented Fine-grained Fusion for Large Language Models (2026.findings-eacl)

Copied to clipboard

Himanshu Chaudhary, Ruida Wang, Gowtham Ramesh, Junjie Hu

Structured Data in NLP

Challenge:	Existing methods to integrate graphs into LLMs compress the graph's structural information into a single token, restricting their ability to capture deep semantic and structural information.
Approach:	They propose a method that integrates fine-grained node-level structural information with corresponding text entities to LLMs via a lightweight, structure adapter module.
Outcome:	The proposed method outperforms baseline models in graph-based question answering by 10.24%.

LoRE: Enhancing Search Relevance with Progressive Chain-of-Thought and Preference Alignment (2026.findings-acl)

Copied to clipboard

Chenji Lu, Zhuo Chen, Hui Zhao, Zhiyuan Zeng, Gang Zhao, Junjie Ren, null Lihaoran, Songyan Liu, Pengjie Wang, Chuan Yu, Jian Xu, Bo Zheng

compare-mt: A Tool for Holistic Comparison of Language Generation Systems (N19-4)

Challenge:	E-commerce search relevance is a critical component of retrieval systems.
Approach:	They propose a large-generative model for search relevance that trains reasoning knowledge, multi-modal understanding and rule awareness into three core competencies.
Outcome:	The proposed model outperforms GPT-5 in Macro-F1 and achieves 27% online gain.

Copied to clipboard

Graham Neubig, Zi-Yi Dou, Junjie Hu, Paul Michel, Danish Pruthi, Xinyi Wang

Text Generation

Challenge:	Unlike machine translation, natural language outputs are nuanced and there are no clear yes/no distinctions about whether they are correct or not.
Approach:	They describe compare-mt, a tool for holistic analysis and comparison of the results of systems for language generation tasks such as machine translation.
Outcome:	The compare-mt tool is an open-source pure-python package that has already proven useful to generate analyses that have been used in our papers.

LLM can Achieve Self-Regulation via Hyperparameter Aware Generation (2024.findings-acl)

Copied to clipboard

Siyin Wang, Shimin Li, Tianxiang Sun, Jinlan Fu, Qinyuan Cheng, Jiasheng Ye, Junjie Ye, Xipeng Qiu, Xuanjing Huang

Text Generation

Challenge:	Existing decoding strategies and hyperparameters may not be optimal for each sample.
Approach:	They propose a model that auto-regulates decoding strategies and hyperparameters . this approach eliminates the need for extensive manual tuning, they argue .
Outcome:	The proposed model eliminates the need for extensive manual tuning, offering a more autonomous, self-regulate model behavior.

How do Role Models Shape Collective Morality? Exemplar-Driven Moral Learning in Multi-Agent Simulation (2026.acl-long)

Copied to clipboard

Junjie Liao, Huacong Tang, Zhou Ziheng, Yizhou Wang, Fangwei Zhong

Linguistics & Cognitive NLP

Challenge:	Existing studies show that role models influence morality, but they are not uniformly interpreted and appropriated in groups with heterogeneous motivations.
Approach:	They build a multi-agent simulation where agents with diverse intrinsic drives interact and adapt through a four-stage cognitive loop.
Outcome:	The proposed model can significantly reshape morality of agents with diverse intrinsic drives . the simulations show that identity-driven conformity can substantially reshaped initial dispositions .

GUICourse: From General Vision Language Model to Versatile GUI Agent (2025.acl-long)

Copied to clipboard

Wentong Chen, Junbo Cui, Jinyi Hu, Yujia Qin, Junjie Fang, Yue Zhao, Chongyi Wang, Jun Liu, Guirong Chen, Yupeng Huo, Yuan Yao, Yankai Lin, Zhiyuan Liu, Maosong Sun

Visual Data in NLP

Challenge:	Graphical User Interfaces (GUIs) are a pivotal medium for human-computer interaction.
Approach:	They propose a series of datasets for training visual-based GUI agents using general VLMs.
Outcome:	The proposed GUICourse datasets show that even a small-sized GUI agent performs better on GUI tasks.

Vulnerability of Text-to-Image Models to Prompt Template Stealing: A Differential Evolution Approach (2025.findings-acl)

Copied to clipboard

Yurong Wu, Fangwen Mu, Qiuhong Zhang, Jinjing Zhao, Xinrun Xu, Lingrui Mei, Yang Wu, Lin Shi, Junjie Wang, Zhiming Ding, Yiwei Wang

LegalAgentBench: Evaluating LLM Agents in Legal Domain (2025.acl-long)

Challenge:	Prompt trading has emerged as a significant intellectual property concern in recent years, where vendors entice users by showcasing sample images before selling prompt templates that can generate similar images.
Approach:	They propose a prompt-stealing benchmark consisting of 50 templates and 450 images organized into Easy and Hard difficulty levels.
Outcome:	The proposed method outperforms baseline methods with an average improvement of over 10%.

Copied to clipboard

Haitao Li, Junjie Chen, Jingli Yang, Qingyao Ai, Wei Jia, Youfeng Liu, Kai Lin, Yueyue Wu, Guozhi Yuan, Yiran Hu, Wuyue Wang, Yiqun Liu, Minlie Huang

Challenge:	Existing general-domain benchmarks do not capture complexity of real-world judicial cognition and decision-making.
Approach:	They propose a benchmark specifically designed to evaluate LLM Agents in the legal domain.
Outcome:	The proposed benchmark includes 17 corpora from real-world legal scenarios and provides 37 tools for interacting with external knowledge.

TL-Training: A Task-Feature-Based Framework for Training Large Language Models in Tool Use (2025.findings-emnlp)

Copied to clipboard

Junjie Ye, Yilong Wu, Sixian Li, Yuming Yang, Zhiheng Xi, Tao Gui, Qi Zhang, Xuanjing Huang, Peng Wang, Zhongchao Shi, Jianping Fan, Zhengyin Du

GroupRank: A Groupwise Paradigm for Effective and Efficient Passage Reranking with LLMs (2026.findings-acl)

Challenge:	a new approach to training large language models (LLMs) overlooks task-specific characteristics in tool use, leading to performance bottlenecks.
Approach:	They propose a task-feature-based framework that mitigates the effects of suboptimal training data . they use a dataset to train large-scale LLMs and a reward mechanism tailored to error categories .
Outcome:	The proposed framework matches or surpasses open- and closed-source LLMs in tool-use performance using only 1,217 training data points.

Copied to clipboard

Meixiu Long, Duolin Sun, Dan Yang, Yihan Jiao, Lei Liu, Jiahai Wang, Binbin Hu, Yue Shen, Jie Feng, Zhehao Tan, Junjie Wang, Lianzhen Zhong, Jian Wang, Peng Wei, Jinjie Gu

From Heads to Neurons: Causal Attribution and Steering in Multi-Task Vision–Language Models (2026.findings-acl)

Challenge:	Existing rankers excel in lexical-matching scenarios, while they struggle with complex queries requiring deep reasoning.
Approach:	They propose a new paradigm that balances flexibility and context awareness to unlock the full potential of groupwise reranking.
Outcome:	The proposed approach achieves a state-of-the-art 65.2 NDCG@10 on BRIGHT and surpasses baselines by 2.1 points on R2MED while delivering a 6.4 inference speedup.

Copied to clipboard

Qidong Wang, Junjie Hu, Ming Jiang

UniEX: An Effective and Efficient Framework for Unified Information Extraction via a Span-extractive Perspective (2023.acl-long)

Challenge:	Existing models focus on single tasks, limiting comparability of neuron importance . ranking strategies overlook how task-dependent information pathways shape write-in effects of feed-forward network (FFN) neurons.
Approach:	They propose a gradient-free framework for task-aware neuron attribution and steering in multi-task vision-language models.
Outcome:	The proposed framework outperforms existing methods in identifying task-critical neurons and improves model performance after steering.

Copied to clipboard

Yang Ping, JunYu Lu, Ruyi Gan, Junjie Wang, Yuxiang Zhang, Pingjian Zhang, Jiaxing Zhang

Seeing the Whole Elephant: A Benchmark for Failure Attribution in LLM-based Multi-Agent Systems (2026.acl-long)

Challenge:	Existing approaches for information extraction (IE) are limited by the number of subtasks and the isolation of the subtask.
Approach:	They propose a new paradigm for universal information extraction that is compatible with any schema format and applicable to a list of IE tasks.
Outcome:	The proposed framework outperforms generative universal IE models on 14 benchmarks with the supervised setting and the state-of-the-art performance in low-resource scenarios.

Copied to clipboard

Mengzhuo Chen, Junjie Wang, Fangwen Mu, Yawen Wang, Zhe Liu, Huanxiang Feng, Qing Wang

One Shot Dominance: Knowledge Poisoning Attack on Retrieval-Augmented Generation Systems (2025.findings-emnlp)

Challenge:	Existing benchmarks rely on partially observable traces that capture only agent outputs . lack of full execution traces obscures many failure causes, authors argue .
Approach:	They propose a benchmark that allows attribution under full execution observability . they find full traces improve attribution accuracy by up to 76.5% over a partial-observation counterpart .
Outcome:	The proposed benchmark improves attribution accuracy by up to 76.5% over a partial-observation counterpart.

Copied to clipboard

Zhiyuan Chang, Mingyang Li, Xiaojun Jia, Junjie Wang, Yuekai Huang, Ziyou Jiang, Yang Liu, Qing Wang

ProFit: Leveraging High-Value Signals in SFT via Probability-Guided Token Selection (2026.findings-acl)

Challenge:	Existing knowledge poisoning attacks against RAG systems require multiple poisoned documents or can only function effectively on simplistic queries.
Approach:	They propose a more realistic knowledge poisoning attack that poisons only a single document while remaining effective for complex multi-hop questions involving complex relationships between multiple elements.
Outcome:	The proposed attack achieves success by poisoning only a single document while remaining effective for complex multi-hop questions involving complex relationships between multiple elements.

Copied to clipboard

Tao Liu, Taiqiang Wu, Runming Yang, Shaoning Sun, Junjie Wang, Yujiu Yang

Challenge:	Traditional fine-tuning ignores one-to-many nature of language, leading to overfitting . authors propose a method to fine- tune LLMs by leveraging tokens.
Approach:	They propose a method to fine-tune Large Language Models by leveraging tokens to mask low-probability tokens.
Outcome:	The proposed method outperforms baselines on general reasoning and mathematical benchmarks.

V-SEAM: Visual Semantic Editing and Attention Modulating for Causal Interpretability of Vision-Language Models (2025.emnlp-main)

Copied to clipboard

Qidong Wang, Junjie Hu, Ming Jiang

Explainability & Interpretability in NLP

Challenge:	Existing work on causal interpretability focuses on large language models (LLMs) but internal mechanisms of vision-language models remain underexplored, authors say .
Approach:	They introduce a framework that combines visual and semantic manipulations for causal interpretation of vision-language models.
Outcome:	The proposed framework shows improved performance for LLAVA and InstructBLIP on three diverse benchmarks.

All Changes May Have Invariant Principles: Improving Ever-Shifting Harmful Meme Detection via Design Concept Reproduction (2026.acl-long)

Copied to clipboard

Ziyou Jiang, Mingyang Li, Junjie Wang, Yuekai Huang, Jie Huang, Zhiyuan Chang, Zhaoyang Li, Qing Wang

Know Thy Enemy: Securing LLMs Against Prompt Injection via Diverse Data Synthesis and Instruction-Level Chain-of-Thought Learning (2026.findings-acl)

Challenge:	Existing methods for harmful meme detection only learn the combination of harmful elements and lack understanding of these implicit expressions.
Approach:	They propose a method that detects harmful memes by replicating the design concept of malicious users.
Outcome:	The proposed method achieves the highest accuracy with 81.1% and has slight accuracy decreases when generalized to type-shifting and temporal-evolving memes.

Copied to clipboard

Zhiyuan Chang, Mingyang Li, Yuekai Huang, Ziyou Jiang, Xiaojun Jia, Qian Xiong, Junjie Wang, Zhaoyang Li, Qing Wang

Multimodal Transformers are Hierarchical Modal-wise Heterogeneous Graphs (2025.acl-long)

Challenge:	Large language model (LLM)-integrated applications face security vulnerabilities from prompt injection (PI) attacks.
Approach:	They propose a model enhancement method that synthesizes diverse training data and employs instruction-level chain-of-thought fine-tuning to enable LLMs to effectively identify and reject malicious instructions regardless of their source or position in the context.
Outcome:	The proposed method outperforms baselines in three critical dimensions while maintaining utility performance without degradation.

Copied to clipboard

Yijie Jin, Junjie Peng, Xuanchao Lin, Haochen Yuan, Lan Wang, Cangzhi Zheng

Entity Relation Extraction as Dependency Parsing in Visually Rich Documents (2021.emnlp-main)

Challenge:	Multimodal Sentiment Analysis (MSA) is a rapidly developing field that integrates multimodal information to recognize sentiments.
Approach:	They propose a multimodal fusion model that integrates multimodal information to recognize sentiments using multimodal transformers.
Outcome:	The proposed model achieves significantly higher performance than MulTs and the existing model is robust.

Copied to clipboard

Yue Zhang, Zhang Bo, Rui Wang, Junjie Cao, Chen Li, Zuyi Bao

Syntactic Text Processing

Challenge:	Existing studies on key information extraction from visually rich documents focus on labeling the text within bounding boxes, while relations between words are unexplored.
Approach:	They propose to use a dependency parsing model to extract semantic entities from visually rich documents by combining entity labeling and relation extraction tasks.
Outcome:	The proposed model achieves 65.96% F1 score on the FUNSD dataset.

Thinking Traps in Long Chain-of-Thought: A Measurable Study and Trap-Aware Adaptive Restart (2026.findings-acl)

Copied to clipboard

null Chenkang, Fan Yu, Junjie Nian, Sihan Zhao, Zhuoka Feng, Zijun Yao, Wang Heng, Yu Minshen, Yixin Cao

Where Did It Go Wrong? Capability-Oriented Failure Attribution for Vision-and-Language Navigation Agents (2026.findings-acl)

Challenge:	Experiments show that extended generation does not guarantee correctness . a recurring pattern in Long-CoT failures is a problem for large reasoning models .
Approach:	They propose a test-time control framework that truncates the trajectory before the trap segment and adaptively restarts decoding.
Outcome:	Experiments show that TAAR improves reasoning performance without fine-tuning model parameters.

Copied to clipboard

Jianming Chen, Yawen Wang, Junjie Wang, Xiaofei Xie, Shoubin Li, Qing Wang, Fanjiang Xu

SAGE: Synergistic Adaptive Gating of Experts for Hateful Video Detection (2026.acl-long)

Challenge:	Existing testing methods are system-level and provide limited insight into which capability deficiencies cause task failures.
Approach:	They propose a capability-oriented testing approach that enables failure detection and attribution by seed selection and mutation.
Outcome:	The proposed method detects more failure cases and pinpoints capability-level deficiencies than state-of-the-art baselines, providing more interpretable and actionable guidance for improving embodied agents.

Copied to clipboard

Jie Huang, Xin Liao, Junjie Wang, Mingyang Li, Wenshuo Wang, Ziyou Jiang, Shoubin Li, Qing Wang

Visual Data in NLP

Challenge:	Existing methods for hateful video detection rely on multimodal feature fusion . existing methods rely only on blind feature mixing, which leads to feature dilution .
Approach:	They propose a framework that shifts from blind feature mixing to decision-level arbitration . it instantiates disentangled experts to rigorously preserve modality-specific semantics .
Outcome:	The proposed framework outperforms state-of-the-art methods on HateMM and MultiHateClip benchmarks.

Attribute-aware Sequence Network for Review Summarization (D19-1)

Copied to clipboard

Junjie Li, Xuepeng Wang, Dawei Yin, Chengqing Zong

Summarization

Challenge:	Existing review summarization systems generate summary only based on review content and neglect the authors’ attributes (e.g., gender, age, and occupation).
Approach:	They propose an Attribute-aware Sequence Network (ASN) to take the aforementioned users’ characteristics into account by encoding their attributes over the words.
Outcome:	The proposed model outperforms existing systems on tripAtt and human evaluation by taking the authors' attributes into account and incorporating attribute embedding and word-using habits into word prediction.

StepCoder: Improving Code Generation with Reinforcement Learning from Compiler Feedback (2024.acl-long)

Copied to clipboard

Shihan Dou, Yan Liu, Haoxiang Jia, Enyu Zhou, Limao Xiong, Junjie Shan, Caishuang Huang, Xiao Wang, Xiaoran Fan, Zhiheng Xi, Yuhao Zhou, Tao Ji, Rui Zheng, Qi Zhang, Tao Gui, Xuanjing Huang

Programming Languages in NLP

Challenge:	Existing work integrates reinforcement learning with compiler feedback to enhance code generation quality but the long code generated by LLMs makes RL exploration ineffective.
Approach:	They propose a framework that integrates reinforcement learning and compiler feedback to enhance code generation quality.
Outcome:	The proposed framework outperforms state-of-the-art approaches in corresponding benchmarks and integrates reinforcement learning with compiler feedback to improve code generation quality.

Solving Math Word Problems via Cooperative Reasoning induced Language Models (2023.acl-long)

Copied to clipboard

Xinyu Zhu, Junjie Wang, Lin Zhang, Yuxiang Zhang, Yongfeng Huang, Ruyi Gan, Jiaxing Zhang, Yujiu Yang

Towards Scalable Lightweight GUI Agents via Multi-role Orchestration (2026.findings-acl)

Challenge:	Large-scale pre-trained language models (PLMs) can be used to solve math word problems, but they lack fast adaptivity as humans.
Approach:	They propose a cooperative reasoning-induced PLM for solving the math word problem . they use system 1 as the generator and system 2 as the verifier to generate reasoning paths .
Outcome:	The proposed model improves on several mathematical reasoning datasets and achieves 9.6% improvement over baselines.

Copied to clipboard

Ziwei Wang, Junjie Zheng, Leyang Yang, Sheng Zhou, Xiaoxuan Tang, Fang Zhouhua, Zhiwei Liu, Dajun Chen, Yong Li, Jiajun Bu

Challenge:	Advanced GUI agents suffer from prohibitive deployment costs on resource-constrained devices.
Approach:	They propose a lightweight GUI agent with GUI-specific knowledge and task scalability . LAMO-3B supports monolithic execution and MAS-style orchestration .
Outcome:	The proposed GUI agent LAMO-3B supports monolithic execution and MAS-style orchestration.

LaERC-S: Improving LLM-based Emotion Recognition in Conversation with Speaker Characteristics (2025.coling-main)

Copied to clipboard

Yumeng Fu, Junjie Wu, Zhongjie Wang, Meishan Zhang, Lili Shan, Yulin Wu, Bingquan Liu

Sentiment Analysis

Challenge:	Emotion recognition in conversation (ERC) is a task of discerning human emotions for each utterance within a conversation.
Approach:	They propose a framework that uses large language models to analyze speaker characteristics . they use two-stage learning to make the models reason speaker characteristics and track emotion of the speaker .
Outcome:	The proposed framework outperforms existing methods on three benchmark datasets.

Analyzing the Effects of Supervised Fine-Tuning on Model Knowledge from Token and Parameter Levels (2025.emnlp-main)

Copied to clipboard

Junjie Ye, Yuming Yang, Yang Nan, Shuo Li, Qi Zhang, Tao Gui, Xuanjing Huang, Peng Wang, Zhongchao Shi, Jianping Fan

VRPO: Rethinking Value Modeling for Robust RL under Noisy Supervision in LLM Post-Training (2026.acl-long)

Challenge:	Large language models (LLMs) acquire substantial world knowledge during pretraining, which is further shaped by post-training techniques such as supervised fine-tuning (SFT).
Approach:	They evaluate closed-book question answering (CBQA) performance across five LLMs from the LLaMA-2 and LLama-3 families and examine the impact of supervised fine-tuning on model knowledge.
Outcome:	The proposed model performance is 14% worse than models fine-tuned on 1,920 samples and 12% worse on 240 samples.

Copied to clipboard

Dingwei Zhu, Shihan Dou, Zhiheng Xi, Senjie Jin, Guoqiang Zhang, Jiazheng Zhang, Junjie Ye, Mingxu Chai, Enyu Zhou, Ming Zhang, Yuhui Wang, Caishuang Huang, Chenhao Huang, Yunke Zhang, Yuran Wang, Tao Gui, Qi Zhang, Xipeng Qiu, Xuanjing Huang

Chain-of-Reasoning: Towards Unified Mathematical Reasoning in Large Language Models via a Multi-Paradigm Perspective (2025.acl-long)

Challenge:	Reinforcement Learning (RL) in real-world environments often suffers from ambiguous or incomplete supervision.
Approach:	They propose a framework that enhances value modeling for robust RL in LLM post-training by integrating auxiliary losses guided by entropy and perplexity from a frozen language model and variational information bottleneck.
Outcome:	The proposed framework outperforms baselines on multi-turn dialogue, math reasoning, and science QA with rule-based and model-based rewards.

Copied to clipboard

Yiyao Yu, Yuxiang Zhang, Dongdong Zhang, Xiao Liang, Hengyuan Zhang, Xingxing Zhang, Mahmoud Khademi, Hany Hassan Awadalla, Junjie Wang, Yujiu Yang, Furu Wei

APEX: Learning Adaptive Priorities for Multi-Objective Alignment in Vision-Language Generation (2026.findings-acl)

Challenge:	Existing work shows that LLMs rely on single-paradigm reasoning that limits their effectiveness across diverse tasks.
Approach:	They propose a new framework that integrates multiple reasoning paradigms to enable synergistic collaboration.
Outcome:	The proposed model outperforms current SOTA models in theorem proving tasks and the MATH benchmark in arithmetic tasks.

Copied to clipboard

Dongliang Chen, Xinlin Zhuang, Junjie Xu, Luojian Xie, Zehui Wang, Jiaxi Zhuang, Haolin Yang, Liang Dou, Xiao He, Xingjiao Wu, Ying Qian

SimpleDeepSearcher: Deep Information Seeking via Web-Powered Reasoning Trajectory Synthesis (2025.findings-emnlp)

Challenge:	APEX optimizes for text-to-image generation by combining learning potential, conflict penalty, and progress need.
Approach:	They propose an algorithm that stabilizes heterogeneous rewards and dynamically schedules objectives . they propose a method that achieves better Pareto trade-offs across four heterogenous objectives based on P3 Adaptive Priorities .
Outcome:	The proposed algorithm achieves better pareto trade-offs across four heterogeneous objectives while maintaining competitive OCR accuracy.

Copied to clipboard

Shuang Sun, Huatong Song, Yuhao Wang, Ruiyang Ren, Jinhao Jiang, Junjie Zhang, Fei Bai, Jia Deng, Xin Zhao, Zheng Liu, Lei Fang, Zhongyuan Wang, Ji-Rong Wen

ToolHop: A Query-Driven Benchmark for Evaluating Large Language Models in Multi-Hop Tool Use (2025.acl-long)

Challenge:	Existing approaches to deep search training lack high-quality training trajectories, prohibitive computational costs and lack of high-fidelity training data.
Approach:	They propose a framework that synthesizes high-quality training data by simulating real user interactions in live web search environments.
Outcome:	The proposed framework synthesizes high-quality training data by simulating user interactions in live web search environments.

Copied to clipboard

Junjie Ye, Zhengyin Du, Xuesong Yao, Weijian Lin, Yufei Xu, Zehui Chen, Zaiyuan Wang, Sining Zhu, Zhiheng Xi, Siyu Yuan, Tao Gui, Qi Zhang, Xuanjing Huang, Jiecao Chen

Concise and Organized Perception Facilitates Reasoning in Large Language Models (2025.findings-naacl)

Challenge:	Effective evaluation of multi-hop tool use is critical for analyzing the understanding, reasoning, and function-calling capabilities of large language models.
Approach:	They propose a dataset that provides rigorous evaluation of multi-hop tool use.
Outcome:	The proposed model achieves 49.04% accuracy across five model families.

Copied to clipboard

Junjie Liu, Shaotian Yan, Chen Shen, Zhengdong Xiao, Liang Xie, Wenxiao Wang, Jieping Ye

A.S.E: A Repository-Level Benchmark for Evaluating Security in AI-Generated Code (2026.findings-acl)

Challenge:	Extensive experimental results on several popular logical benchmarks (ProofWriter, PrOntoQA, PrONtoQA-OOD, and FOLIO) and mathematical benchmark (DI-GSM) show that COP significantly outperforms previous state-of-the-art methods.
Approach:	They propose a reasoning approach called Concise and Organized Perception (COP) that carefully analyzes the given statements to identify the most pertinent information while eliminating redundancy efficiently.
Outcome:	The proposed approach outperforms state-of-the-art methods on several popular logical benchmarks and mathematical benchmarks.

Copied to clipboard

Keke Lian, Wang Bin, Lei Zhang, Libo Chen, Junjie Wang, Ziming Zhao, Yujiu Yang, Miaoqian Lin, Haotong Duan, Haoran Zhao, Shuang Liao, Mingda Guo, Quan Jiazheng, Yilu Zhong, Chenhao He, Chen Zichuan, Jie Wu, Haoling Li, Zhaoxuan Li, Jiongchi Yu, Hui LI, Dong Zhang

Programming Languages in NLP

Challenge:	Existing security evaluation benchmarks lack relevance to real-world AI programming tasks . current LLMs struggle with secure coding, research shows .
Approach:	They propose a repository-level evaluation benchmark to assess security of AI-generated code.
Outcome:	The proposed framework mirrors real-world AI programming tasks and offers valuable insights into the state of AI code generation.

HoLLMwood: Unleashing the Creativity of Large Language Models in Screenwriting via Role Playing (2024.findings-emnlp)

Copied to clipboard

Jing Chen, Xinyu Zhu, Cheng Yang, Chufan Shi, Yadong Xi, Yuxiang Zhang, Junjie Wang, Jiashu Pu, Tian Feng, Yujiu Yang, Rongsheng Zhang

OntoGuard: Enforcing Action Admissibility for LLM Agents in Complex Interactive Environments (2026.findings-acl)

Challenge:	Generative AI has demonstrated unprecedented creativity in the field of computer vision, yet such phenomena have not been observed in the realm of literary creation.
Approach:	They propose a framework for unleashing the creativity of large language models (LLMs) they assign LLMs to different roles involved in real-world scenario, they write .
Outcome:	The proposed framework outperforms baselines in terms of coherence, relevance, interestingness and overall quality on automatically generated screenplays.

Copied to clipboard

Pengxiang Liu, Tao Ren, Wei Xiong, Tingrui Yang, Junjie Wang, Jun HU

Challenge:	Existing approaches to large language models (LLMs) are limited by their ability to enforce environmental and behavioral admissibility.
Approach:	They propose an ontological framework to guard LLM agents by enforcing environmental and behavioral admissibility.
Outcome:	Experiments on ScienceWorld and VirtualHome show that OntoGuard can enforce environmental and behavioral admissibility while preventing invalid actions.

Towards Accurate and Consistent Evaluation: A Dataset for Distantly-Supervised Relation Extraction (2020.coling-main)

Copied to clipboard

Tong Zhu, Haitao Wang, Junjie Yu, Xiabing Zhou, Wenliang Chen, Wei Zhang, Min Zhang