Papers by Yan Zeng

Novel Slot Detection: A Benchmark for Discovering Unknown Slot Types in the Task-Oriented Dialogue System (2021.acl-long)

Challenge:	Existing classification and regression models that only extract finer-grained information from magnetic resonance imaging (MRI) may not be effective for Alzheimer's disease (AD).
Approach:	They propose to use a 3D Adapter in a Vision Transformer to extract the patient's EHR information and questions related to the disease as text prompts.
Outcome:	The proposed model can discriminate and predict the corresponding MMSE score based on the extracted brain structural information and textual content .

Copied to clipboard

Yanan Wu, Zhiyuan Zeng, Keqing He, Hong Xu, Yuanmeng Yan, Huixing Jiang, Weiran Xu

Natural Language Interfaces

Challenge:	Existing slot filling models can only recognize pre-defined in-domain slot types from a limited slot set.
Approach:	They introduce a task, Novel Slot Detection, in the task-oriented dialogue system.
Outcome:	The proposed task is based on two public NSD datasets and proposes strong baselines . it aims to identify a sequence of tokens and extract semantic constituents from user queries .

Cross-View Language Modeling: Towards Unified Cross-Lingual Cross-Modal Pre-training (2023.acl-long)

Copied to clipboard

Yan Zeng, Wangchunshu Zhou, Ao Luo, Ziming Cheng, Xinsong Zhang

SafeSteer: A Decoding-level Defense Mechanism for Multimodal Large Language Models (2026.findings-acl)

Challenge:	Empirical results show that CCLM significantly outperforms the prior state-of-the-art with an average absolute improvement of over 10%.
Approach:	They introduce a pre-training framework that unifies cross-lingual and cross-modal pre-trained models with shared architectures and objectives.
Outcome:	The proposed framework outperforms the state-of-the-art in two multi-lingual datasets and two multilingual image-text retrieval datasets.

Copied to clipboard

Xinyi Zeng, Xue Yang, Jingyuan Zhang, Huanqian Yan, Xiang Chen, Kaiwen Wei, Hankun Kang, Yu Tian

Responsible & Trustworthy NLP

Challenge:	Existing defense methods rely on fine-tuning or inefficient post-hoc interventions, limiting their ability to address novel attacks.
Approach:	They propose a decoding-level defense mechanism that employs a lightweight discriminator to iteratively steer the decoding process toward safety.
Outcome:	The proposed method improves safety performance by up to 33.40% without fine-tuning on multiple MLLMs.

Beyond Noise: Characterizing Creative Potential in Unverifiable LLM Hallucinations (2026.acl-long)

Copied to clipboard

Yu Yan, Chunhong Zhang, Haiyu Zhao, Ziyang Zeng, Zihao Liu, Yongkang Wu, Jianzhou Diao, YiJie Chen, Shujie Wang, Zheng Hu

Reasoning

Challenge:	Large Language Models generate outputs that extend beyond established knowledge . prior work does not characterize the unverifiable space as a whole .
Approach:	They propose a novelty-verifiability characterization that distinguishes Creative Synthesis from Groundless Fabrication by a conceptual creation task.
Outcome:	The proposed model distinguishes Creative Synthesis (Region A) from Groundless Fabrication (Regium B) it shows that Region A is non-negligible and robust, persisting across generation strategies, models, domains, and embedding choices.

PersLLM: A Personified Training Approach for Large Language Models (2025.findings-emnlp)

Copied to clipboard

Zheni Zeng, Jiayi Chen, Huimin Chen, Yukun Yan, Yuxuan Chen, Zhenghao Liu, Zhiyuan Liu, Maosong Sun

KBAlign: Efficient Self Adaptation on Specific Textual Knowledge Bases (2025.findings-emnlp)

Challenge:	Large language models exhibit human-like intelligence, enabling them to simulate human behavior and support various applications that require both humanized communication and extensive knowledge reserves.
Approach:	They propose a framework for better data construction and model tuning to unlock the potential of LLM personification by using Chain-of-Thought prompting and anti-induction.
Outcome:	The proposed framework improves data construction and model tuning for insufficient data usage and rigid behavior patterns.

Copied to clipboard

Zheni Zeng, Yuxuan Chen, Shi Yu, Ruobing Wang, Yukun Yan, Zhenghao Liu, Shuo Wang, Xu Han, Zhiyuan Liu, Maosong Sun

Judge as A Judge: Improving the Evaluation of Retrieval-Augmented Generation through the Judge-Consistency of Large Language Models (2025.findings-acl)

Challenge:	Existing methods for retrieval-augmented generation (RAG) are limited and fine-tuning incurs prohibitive costs of external signals.
Approach:	They propose a self-supervised framework that enhances RAG systems through efficient model adaptation.
Outcome:	The proposed framework achieves 90% of the performance gain obtained through GPT-4-supervised adaptation while relying entirely on self-annotation of much smaller models.

Copied to clipboard

Shuliang Liu, Xinze Li, Zhenghao Liu, Yukun Yan, Cheng Yang, Zheni Zeng, Zhiyuan Liu, Maosong Sun, Ge Yu

Turn Waste into Worth: Rectifying Top-k Router of MoE (2024.emnlp-main)

Challenge:	Existing evaluation metrics cannot fairly evaluate the outputs of RAG models during training and evaluation.
Approach:	They propose a method which prompts LLMs to generate different judgments based on various combinations of judgment dimensions and utilizes the judge-consistency to evaluate these judgments.
Outcome:	The proposed method generates more accurate evaluations for RAG models across different RAG model and datasets.

Copied to clipboard

Zhiyuan Zeng, Qipeng Guo, Zhaoye Fei, Zhangyue Yin, Yunhua Zhou, Linyang Li, Tianxiang Sun, Hang Yan, Dahua Lin, Xipeng Qiu

Green & Sustainable NLP

Challenge:	Top-k router suffers from redundancy computation and memory costs due to unbalanced routing . some experts are overflow, where exceeding tokens are dropped, while others are empty, which are padded with zeros, negatively impacting model performance.
Approach:	They propose a top-k router that is unbalanced and uses a multi-gPU system to handle dropped tokens and padding.
Outcome:	The proposed model surpasses the top-1 router by 4.7% in terms of performance . the top-k router suffers from redundancy computation and memory costs .

Toward Building General Foundation Models for Language, Vision, and Vision-Language Understanding Tasks (2023.findings-emnlp)

Copied to clipboard

Xinsong Zhang, Yan Zeng, Jipeng Zhang, Hang Li

Stepwise Perplexity-Guided Refinement for Efficient Chain-of-Thought Reasoning in Large Language Models (2025.findings-acl)

Challenge:	Existing foundation models can only perform the best in one type of understanding tasks.
Approach:	They propose a method for training a general foundation model, X-FM, using text, image, and image-text data.
Outcome:	The proposed method outperforms existing foundation models on language, vision, and vision-language understanding tasks.

Copied to clipboard

Yingqian Cui, Pengfei He, Jingying Zeng, Hui Liu, Xianfeng Tang, Zhenwei Dai, Yan Han, Chen Luo, Jing Huang, Zhen Li, Suhang Wang, Yue Xing, Jiliang Tang, Qi He

Harnessing Consistency for Robust Test-Time LLM Ensemble (2026.findings-eacl)

Challenge:	Chain-of-Thought (CoT) reasoning has improved the performance of large language models (LLMs) however, the detailed reasoning process in CoT often incurs long generation times and high computational costs due to the inclusion of unnecessary steps.
Approach:	They propose a method to identify critical reasoning steps using perplexity as a measure of their importance.
Outcome:	The proposed method achieves a better balance between reasoning accuracy and efficiency of CoT.

Copied to clipboard

Zhichen Zeng, Qi Yu, Xiao Lin, Ruizhong Qiu, Xuying Ning, Tianxin Wei, Yuchen Yan, Jingrui He, Hanghang Tong

Robustness in NLP

Challenge:	Existing efforts to improve LLM ensemble quality have focused on model consistency, but failures are often due to heterogeneous tokenization schemes and varying model expertise.
Approach:	They propose a plug-and-play technique that harnesses model consistency for robust LLM ensemble.
Outcome:	The proposed technique improves ensemble performance and robustness against erroneous signals.

Modeling Discriminative Representations for Out-of-Domain Detection with Supervised Contrastive Learning (2021.acl-short)

Copied to clipboard

Zhiyuan Zeng, Keqing He, Yuanmeng Yan, Zijun Liu, Yanan Wu, Hong Xu, Huixing Jiang, Weiran Xu

Reinforcement Learning on Pre-Training Data (2026.acl-long)

Challenge:	Existing methods of OOD detection only focus on whether a sample is correctly classified . lack of real OOD examples leads to poor prior knowledge about these unknown intents .
Approach:	They propose a supervised contrastive learning objective to minimize intra-class variance . they employ an adversarial augmentation mechanism to obtain pseudo diverse views .
Outcome:	The proposed method minimizes intra-class variance by pulling together in-domain intents belonging to the same class and maximizes inter-class variation by pushing apart samples from different classes.

Copied to clipboard

Siheng Li, Kejiao Li, Zenan Xu, Guanhua Huang, Kun Li, Haoyuan Wu, null Wujiajia, Zihao Zheng, Chenchen Zhang, Kun Shi, Xue Gong, Qi Yi, Ruibin Xiong, Tingqiang Xu, Yuhao Jiang, Jianfeng Yan, Yuyuan Zeng, Guanghui Xu, Jinbao Xue, Zhijiang xu, Zheng Fang, Shuai LI, Qibin Liu, Xiaoxue Li, Zhuoyu Li, Yangyu Tao, Fei Gao, Cheng Jiang, Bochao Wang, Kai Liu, Jianchen Zhu, Wai Lam, Bo Zhou, Di Wang

Beyond Static Persona Consistency: Dynamic Persona Coherence in LLM Role-Playing (2026.acl-long)

Challenge:	Recent progress in large language models is driven by scaling of training compute through pre-training with nexttoken prediction (NTP) or post-training (RL) Pre-training using NTP enables models to acquire extensive knowledge and skills from general data, but it suffers from data inefficiency and catastrophic forgetting in continual learning settings.
Approach:	They propose to scale training compute through pre-training with next-token prediction (NTP) or post-training by scaling reinforcement learning (RL) to improve learning from general data.
Outcome:	Experiments on multiple benchmarks and models show that the proposed approach improves continual pre-training and provides a strong foundation for post-training on Qwen3-8B-Base.

Copied to clipboard

Yirui QI, Xiaoming Zhang, Ruilin Zeng, Mengyao Liu, Ziyi Zhou, Dezhuang Miao, Bingyu Yan, Zhenyu Guan

Challenge:	Existing LLMs conflate identity consistency with emotional rigidity . Existing models exhibit either robotic repetition or persona drift .
Approach:	They propose a framework that decouples Identity-Layer Stability from Adaptive-Layer Appropriateness to achieve persona coherence repair.
Outcome:	Experiments on GPT-4o, Claude-3.5-Sonnet, and DeepSeek-V3.2 show consistent improvements (+16–84% gains)

ThinkNote: Enhancing Knowledge Integration and Utilization of Large Language Models via Constructivist Cognition Modeling (2026.findings-eacl)

Copied to clipboard

Zhipeng Xu, Zhenghao Liu, Yukun Yan, Shuo Wang, Shi Yu, Zheni Zeng, Chaojun Xiao, Zhiyuan Liu, Ge Yu, Chenyan Xiong

What Matters in Training a GPT4-Style Language Model with Multimodal Inputs? (2024.naacl-long)

Challenge:	Large Language Models (LLMs) exhibit suboptimal behaviors and inconsistencies when exposed to unfamiliar external information, underscoring their limitations in effectively leveraging such knowledge.
Approach:	They propose a framework that enhances the external knowledge utilization of Large Language Models through a two-stage constructivist cognitive modeling process.
Outcome:	The proposed framework achieves a 10% improvement over baseline methods on various question-answering benchmarks.

Copied to clipboard

Yan Zeng, Hanbo Zhang, Jiani Zheng, Jiangnan Xia, Guoqiang Wei, Yang Wei, Yuchen Zhang, Tao Kong, Ruihua Song

LexicalAT: Lexical-Based Adversarial Reinforcement Training for Robust Sentiment Classification (D19-1)

Challenge:	Recent advances in GPT-4V have demonstrated remarkable multi-modal capabilities in processing image inputs and following open-ended instructions.
Approach:	They propose a plug-and-play technique to enhance multi-modal LLMs . they propose 'lynx' to train multi-modal LLM models .
Outcome:	The proposed training strategy improves understanding accuracy and instruction-following proficiency of multi-modal models.

Copied to clipboard

Jingjing Xu, Liang Zhao, Hanqi Yan, Qi Zeng, Yun Liang, Xu Sun

Robustness in NLP

Challenge:	Existing text classification models are fragile and sensitive to simple perturbations.
Approach:	They propose a generator-classifier adversarial training approach to improve classification models . they use a large-scale lexical knowledge base to generate attacking examples .
Outcome:	The proposed approach outperforms strong baselines and reduces test errors on neural networks.

RankCoT: Refining Knowledge for Retrieval-Augmented Generation through Ranking Chain-of-Thoughts (2025.acl-long)

Copied to clipboard

Mingyan Wu, Zhenghao Liu, Yukun Yan, Xinze Li, Shi Yu, Zheni Zeng, Yu Gu, Ge Yu

Text Generation

Challenge:	Retrieval-Augmented Generation (RAG) models enable Large Language Models to access external knowledge.
Approach:	They propose a knowledge refinement method that incorporates reranking signals to generate CoT-based summarization based on query and retrieval documents.
Outcome:	RankCoT generates CoT-based summarization based on query and all retrieval documents . Rank CoT incorporates a self-reflection mechanism that refines the outputs .

HighMATH: Evaluating Math Reasoning of Large Language Models in Breadth and Depth (2025.findings-emnlp)

Copied to clipboard

Yan Liu, Minghui Zhang, Bojian Xiong, Yifan Xiao, Yinong Sun, Yating Mei, Longyu Zeng, Jingchao Yang, Yang Wang, Deyi Xiong

Reasoning

Challenge:	a gap in math models' accuracy has been widened with the development of large language models (LLMs) . a new study aims to bridge this gap by evaluating a set of high-level math reasoning models .
Approach:	They propose to evaluate large language models on existing math benchmarks to bridge this gap . they collect 5,293 problems from Chinese senior high school mathematics exams .
Outcome:	The proposed model is based on o1-like models and a high-level model.

A Functionality-Grounded Benchmark for Evaluating Web Agents in E-commerce Domains (2026.acl-long)

Copied to clipboard

Xianren Zhang, Shreyas Prasad, Di Wang, Qiuhai Zeng, Suhang Wang, Wenbo Yan, Mat Hans

Challenge:	Existing benchmarks focus on product search tasks, but ignore potential risks.
Approach:	They propose a data generation pipeline that leverages webpage content and interactive elements to create diverse, functionality-grounded user queries.
Outcome:	The proposed framework assesses the performance and safety of web agents under dynamic, real-world e-commerce environments.

Me-Agent: A Personalized Mobile Agent with Two-Level User Habit Learning for Enhanced Interaction (2026.findings-acl)

Copied to clipboard

Shuoxin Wang, Chang Liu, Gowen Loo, Lifan Zheng, Kaiwen Wei, Huanqian Yan, Xinyi Zeng, Jingyuan Zhang, Yu Tian

A Simple and Efficient Multi-Task Learning Approach for Conditioned Dialogue Generation (2021.naacl-main)

Challenge:	Existing Large Language Model (LLM)-based mobile agents follow explicit user instructions without personalized needs.
Approach:	They propose a user preference learning strategy enhanced with a Personal Reward Model to improve personalization performance.
Outcome:	The proposed agent achieves state-of-the-art performance while maintaining competitive instruction execution performance.

Copied to clipboard

Yan Zeng, Jian-Yun Nie

Responsible & Trustworthy NLP

Challenge:	Existing studies have focused on conditioned dialogue generation, but there is a scarcity of labeled responses.
Approach:	They propose a multi-task learning approach to leverage labeled dialogue and text data to generate conditioned dialogues.
Outcome:	The proposed approach outperforms the state-of-the-art models by leveraging the labeled texts and obtains larger improvement compared to the previous methods to leverage text data.

Fin-STAR: Structure-as-Semantics to Resolve Implicitness in Financial Retrieval (2026.findings-acl)

Copied to clipboard

Yu Zou, Yan Chen, Lida He, Qi Zhou, Xiaorui Zhou, Aixi Zhong, Yi Wang, Wei Li, Qingyu Wang, Jiatao Li, Wei Gong, Jialei Zeng, Jingmei Zhao, Ke Jiang, Qing Li

Information Retrieval

Challenge:	Existing Retrieval-Augmented Generation systems treat structure as a physical navigational skeleton rather than intrinsic semantic knowledge.
Approach:	They propose a framework that redefining hierarchy as intrinsic semantics and uses snippets to enrich hierarchical lineage.
Outcome:	The proposed framework outperforms state-of-the-art hierarchical and graph-based benchmarks on FinTierQA Gold.

Detecting AI-Generated Content on Social Media with Multi-modal Language Models (2026.acl-industry)

Copied to clipboard

Chenyang Yang, Shen Yan, Yibo Yang, Litao Hu, Yuchen Liu, Yuan Zeng, Hanchao Yu, Yinan Zhu, Sumedha Singla, Brian Vanover, Huijun Qian, Zihao Wang, Fujun Liu, Aashu Singh, Jianyu Wang, Xuewen Zhang

Topic Memory Networks for Short Text Classification (D18-1)

Challenge:	Existing methods for AI-generated content detection face poor generalization to newer models, reliance on single modalities, and lack of interpretable explanations.
Approach:	They propose a model that curates diverse social media data and trains a vision-language model for detection and explanation.
Outcome:	The proposed model achieves state-of-the-art detection performance on public benchmarks and observes positive downstream impacts on user engagement.

Copied to clipboard

Jichuan Zeng, Jing Li, Yan Song, Cuiyun Gao, Michael R. Lyu, Irwin King

Information Extraction & Text Mining

Challenge:	Existing classification models for short texts are weak due to data sparsity .
Approach:	They propose topic memory networks for short text classification with a novel topic memory mechanism to encode latent topic representations indicative of class labels.
Outcome:	The proposed model outperforms state-of-the-art models on short text classification, while generating coherent topics.

Revisit Overconfidence for OOD Detection: Reassigned Contrastive Learning with Adaptive Class-dependent Threshold (2022.naacl-main)

Copied to clipboard

Yanan Wu, Keqing He, Yuanmeng Yan, QiXiang Gao, Zhiyuan Zeng, Fujia Zheng, Lulu Zhao, Huixing Jiang, Wei Wu, Weiran Xu

CharTide: Data-Centric Chart-to-Code Generation via Tri-Perspective Tuning and Inquiry-Driven Evolution (2026.acl-long)

Challenge:	Existing methods for OOD detection are based on labeled in-domain data . detecting out-of-domain (OOD) or unknown intents is challenging .
Approach:	They propose a novel reassigned contrastive learning method to discriminate IND intents for over-confident OOD and an adaptive class-dependent local threshold mechanism to separate similar IND and OOD intents.
Outcome:	The proposed method is effective for both aspects of overconfidence issues.

Copied to clipboard

Xiangxi Zheng, Kuang He, Jiayi Hu, Ping Yu, Rui Yan, Yuan Yao, Peng Hou, Anxiang Zeng, Alex Jinpeng Wang

MUSE: MCTS-Driven Red Teaming Framework for Enhanced Multi-Turn Dialogue Safety in Large Language Models (2025.emnlp-main)

Challenge:	Existing approaches to chart-to-code generation are constrained by data-centric limitations . authors present a new framework that redesigns both training and alignment data .
Approach:	They propose a data-centric framework that redesigns both training and alignment data for chart-to-code generation.
Outcome:	The proposed framework outperforms open-source baselines and is competitive with GPT-5.

Copied to clipboard

Siyu Yan, Long Zeng, Xuecheng Wu, Chengcheng Han, Kongcheng Zhang, Chong Peng, Xuezhi Cao, Xunliang Cai, Chenjuan Guo

Robustness in NLP

Challenge:	Existing defenses target single-turn attacks, but real-world usage involves multi-turn dialogues, exposing models to attacks that exploit conversational context to bypass safety measures.
Approach:	They propose a framework that tackles multi-turn jailbreaks from both attack and defense angles.
Outcome:	Experiments on large language models show that MUSE effectively mitigates multi-turn jailbreaks.

Adversarial Self-Supervised Learning for Out-of-Domain Detection (2021.naacl-main)

Copied to clipboard

Zhiyuan Zeng, Keqing He, Yuanmeng Yan, Hong Xu, Weiran Xu

Responsible & Trustworthy NLP

Challenge:	Existing methods for detecting out-of-domain (OOD) intents are unsupervised and require extensive labeled data.
Approach:	They propose a self-supervised contrastive learning framework to model discriminative semantic features from unlabeled data.
Outcome:	The proposed framework outperforms baseline methods on two public benchmark datasets with a statistically significant margin.

An Investigation of Suitability of Pre-Trained Language Models for Dialogue Generation – Avoiding Discrepancies (2021.findings-acl)

Copied to clipboard

Yan Zeng, Jian-Yun Nie

Dialogue Systems & Conversational Agents

Challenge:	Pre-trained language models have been widely used in open-domain dialogue generation.
Approach:	They propose to use decoder-only architecture to achieve excellent performance for dialogue generation.
Outcome:	The proposed frameworks are based on transformer-ED, transformer-Dec, transformer MLM and transformer-AR.

SEMIROUTER: Sparse-Data Enhanced Routing for Adaptive Multi-LLM System (2026.eacl-long)

Copied to clipboard

Zijie Wang, Xinyu Yan, Che Wang, Zeng Zihao, Lei Xiao, Wei Yang Bryan Lim

Green & Sustainable NLP

Challenge:	Existing routing methods suffer from poor scalability and dependence on datasets for training . energy footprint is also considered in the decision to implement our new LLM routing framework .
Approach:	They propose a new LLM routing framework that dynamically allocates queries to the most appropriate LLM.
Outcome:	The proposed method improves data efficiency, adaptability, and routing accuracy compared to existing methods.

EfficientVLM: Fast and Accurate Vision-Language Models via Knowledge Distillation and Modal-adaptive Pruning (2023.findings-acl)

Copied to clipboard

Tiannan Wang, Wangchunshu Zhou, Yan Zeng, Xinsong Zhang