Papers by Zhang Yijun
Multiview Clickbait Detection via Jointly Modeling Subjective and Objective Preference (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing clickbait detection models rely on analyzing the objective semantics of posts or correlating posts with article content only, but fail to identify and exploit the manipulation intention of clickbaiting from a user’s subjective perspective. |
| Approach: | They propose a multiview clickbait detection model to model subjective and objective preferences simultaneously to capture clickbaiting from a user's subjective perspective. |
| Outcome: | The proposed model outperforms state-of-the-art models on two real-world datasets and shows that it integrates subjective and objective preferences simultaneously. |
Graph-GRPO: Stabilizing Multi-Agent Topology Learning via Group Relative Policy Optimization (2026.findings-acl)
Copied to clipboard
Yueyang Cang, Xiaoteng Zhang, Erlu Zhao, Zehua Ji, Yuhang Liu, Yuchen He, Zhiyuan Ning, Chen Yijun, Wenge Que, Li Shi
| Challenge: | Recent approaches to optimize communication topology rely on single-sample policy gradients with absolute rewards. |
| Approach: | They propose a topology optimization framework that integrates Group Relative Policy Optimization. |
| Outcome: | The proposed topology optimization framework outperforms state-of-the-art methods on reasoning and code generation benchmarks. |
Make Some Noise: Unlocking Language Model Parallel Inference Capability through Noisy Training (2024.emnlp-main)
Copied to clipboard
Yixuan Wang, Xianzhen Luo, Fuxuan Wei, Yijun Liu, Qingfu Zhu, Xuanyu Zhang, Qing Yang, Dongliang Xu, Wanxiang Che
| Challenge: | Existing speculative decoding methods require additional model structure and training processes to assist the model for draft token generation. |
| Approach: | They propose a make some noise training framework that introduces some noise at the input for the model to learn the denoising task. |
| Outcome: | The proposed model improves inference speed by 2.3-2.7x times without compromising model performance. |
RouteMoA: Dynamic Routing without Pre-Inference Boosts Efficient Mixture-of-Agents (2026.acl-long)
Copied to clipboard
Jize Wang, Han Wu, Zhiyuan You, Yiming Song, Yijun Wang, Zifei Shan, Yining Li, Songyang Zhang, Xinyi Le, Cailian Chen, Xinping Guan, Dacheng Tao
| Challenge: | Existing methods for mixing-of-agents (MoA) lack model selection criteria and struggle with large model pools. |
| Approach: | They propose a mixture-of-agents framework with dynamic routing that uses a lightweight scorer to perform initial screening and refines the model scores through self- and cross-assessment. |
| Outcome: | The proposed framework outperforms existing methods for large model pools and tasks . it reduces cost by 89.8% and latency by 63.6% in the large-scale model pool. |
Domain-aware and Co-adaptive Feature Transformation for Domain Adaption Few-shot Relation Extraction (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing approaches to relation extraction focus on the source domain, which makes it difficult to accurately transfer useful knowledge to the target domain. |
| Approach: | They propose a domain-aware and co-adaptive feature transformation approach to address these issues by leveraging the target domain distribution features to guide the domain-based feature transformations. |
| Outcome: | The proposed method outperforms existing models and achieves state-of-the-art performance on a benchmark dataset. |
On the Perception Bottleneck of VLMs for Chart Understanding (2025.findings-emnlp)
Copied to clipboard
| Challenge: | a perception bottleneck in large vision-language models is critical for chart understanding . instruction tuning improves the extraction capability of LVLMs, but the vision encoder remains a bottleneck . |
| Approach: | They propose to decompose the perception bottleneck into two components . the vision encoder bottleneck is where visual representation fails to encapsulate the correct information . |
| Outcome: | The proposed approach significantly mitigates the vision encoder bottleneck and improves the ability of LVLMs to comprehend charts. |
Chain-of-Talkers (CoTalk): Fast Human Annotation of Dense Image Captions (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches for optimizing human annotation efforts are limited . et al., 2015) suggest that densely annotated image captions improve vision-language alignment . |
| Approach: | They propose an AI-in-the-loop methodology to maximize the number of annotated samples and improve their comprehensiveness under fixed budget constraints. |
| Outcome: | The proposed method improves annotation speed and retrieval performance over the parallel method. |
Adaptive Zooming via Relevance-Informed Positional Resource Allocation for Training-free LLM Context Extension (2026.findings-acl)
Copied to clipboard
Hongbo Zhao, Huibin Wang, Bin Tang, Xianming Hu, Yihong Huang, Yijun Shen, Nuoyi Chen, Ping Li, Kai Zhang
| Challenge: | Existing training-free methods for extrapolating beyond training context lengths are semantics-agnostic . Existing methods that focus on relative token distances can indiscriminately blur semantically relevant and irrelevant tokens . |
| Approach: | They propose an adaptive positional zooming method that uses semantic relevance to extrapolate beyond training context lengths. |
| Outcome: | Experiments show that RiPRA outperforms existing training-free extrapolation methods . relevant tokens get higher positional resolution, while irrelevant tokens are compressed . |
Adversarial Preference Learning for Robust LLM Alignment (2025.findings-acl)
Copied to clipboard
Yuanfu Wang, Pengyu Wang, Chenyang Xi, Bo Tang, Junyi Zhu, Wenqiang Wei, Chen Chen, Chao Yang, Jingfeng Zhang, Chaochao Lu, Yijun Niu, Keming Mao, Zhiyu Li, Feiyu Xiong, Jie Hu, Mingchuan Yang
| Challenge: | Modern language models rely on Reinforcement Learning from Human Feedback (RLHF) to encourage safe behaviors, but they remain vulnerable to adversarial attacks due to three key limitations: (1) the inefficiency and high cost of human annotation; (2) the vast diversity of potential adversarials; and (3) the risk of feedback bias and reward hacking. |
| Approach: | They propose an iterative adversarial training method that incorporates three key innovations to address these challenges. |
| Outcome: | Experiments on Mistral-7B-Instruct-v0.3 show that the proposed method significantly enhances robustness and reduces harmful outputs from 5.88% to 0.43%. |
Escaping the Echo Trap: On Credit Assignment Failure in Multi-turn LLM Self-Reflection (2026.acl-long)
Copied to clipboard
Linxuan Du, Guangquan Xue, Xiaobo Liang, Qipeng Huang, Yuyang Ding, Xinyu Shi, Zhang Yijun, Ji Qi, Wenpeng Zhu, Juntao Li, Min Zhang
| Challenge: | Existing methods for multi-turn self-reflection are limited by the Echo Trap problem . the model is limited by its inherent capabilities and repeats earlier reflections to preserve reward signals . |
| Approach: | They propose a tree-structured extension of GRPO for multi-turn self-reflection which enables more accurate advantage estimation. |
| Outcome: | The proposed method mitigates behavior collapse and improves performance across benchmarks. |
Confusion is the Final Barrier: Rethinking Jailbreak Evaluation and Investigating the Real Misuse Threat of LLMs (2025.findings-emnlp)
Copied to clipboard
Yu Yan, Sheng Sun, Zhe Wang, Yijun Lin, Zenghao Duan, Zhifei Zheng, Min Liu, Zhiyi Yin, Jianping Zhang
| Challenge: | Large Language Models have been developed to deal with real-world crimes, but it remains unclear whether they internalize authentic knowledge or are forced to simulate toxic language patterns. |
| Approach: | They construct knowledge-intensive Q&A to investigate misuse threats of Large Language Models in terms of dangerous knowledge possession, harmful task planning utility, and harmfulness judgment robustness. |
| Outcome: | The findings raise concerns that jailbreak success is often attributable to a hallucination loop between jailbroken LLM and judger LLM . |