Papers by Suhang Wang
TOP-Training: Target-Oriented Pretraining for Medical Extractive Question Answering (2025.coling-main)
Copied to clipboard
| Challenge: | e-health records underscore the growing significance of information extraction (IE) from these datasets. |
| Approach: | They propose a target-oriented pre-training paradigm for extractive question-answering in the medical domain . TOP-Training moves one step further than popular domain-oriented fine-tuning . |
| Outcome: | The proposed method improves on the Medical-EQA benchmarks. |
GAPO: Learning Preferential Prompt through Generative Adversarial Policy Optimization (2025.acl-long)
Copied to clipboard
Zhouhong Gu, Xingzhou Chen, Xiaoran Shi, Tao Wang, Suhang Zheng, Tianyu Li, Hongwei Feng, Yanghua Xiao
| Challenge: | Existing methods for achieving this require a limited understanding of constraints and can be hallucinating or brittle. |
| Approach: | They propose a framework that combines adversarial training dynamics with an encoder-only reward model to progressively learn and adapt to increasingly complex constraints. |
| Outcome: | Extensive experiments show that GAPO significantly outperforms existing methods like PPO, DPO, and KTO in fine-grained constraints. |
Exploring Language Model Generalization in Low-Resource Extractive QA (2025.coling-main)
Copied to clipboard
| Challenge: | Existing LLMs struggle with dataset demands of closed domains such as medicine and law . current LLM performance in closed domain is lacking, even on traditional tasks such as Natural Language Inference . |
| Approach: | They investigate Extractive Question Answering (EQA) with Large Language Models (LLMs) under domain drift . they find that LLMs struggle with dataset demands of closed domains . |
| Outcome: | The proposed model performs poorly in extractive question answering tasks under domain drift . the proposed model can generalize to domains that require specific knowledge without training . |
Query-Efficient Agentic Graph Extraction Attacks on GraphRAG Systems (2026.acl-long)
Copied to clipboard
| Challenge: | Existing attacks exploit leakage of retrieved subgraphs, leaving the security implications of structured knowledge representations unexplored. |
| Approach: | They propose a framework that leverages a novelty-guided exploration–exploitation strategy and external graph memory modules to extract a latent entity–relation graph. |
| Outcome: | The proposed framework outperforms baselines on medical, agriculture, and literary datasets under identical query budgets while maintaining high precision. |
ToolDreamer: Instilling LLM Reasoning Into Tool Retrievers (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing retrieval models rank tools based on similarity between query and tool description (TD) Existing tools are not conditioned to learn tool-to-tool relationships (middle). |
| Approach: | They propose a framework that conditions retrieval models to fetch tools based on hypothetical (synthetic) TD generated using an LLM. |
| Outcome: | The proposed framework improves the performance of sparse and dense retrievers with and without training, showcasing its flexibility. |
Divide-Verify-Refine: Can LLMs Self-align with Complex Instructions? (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing research shows LLMs struggle with complex instructions involving multiple constraints. |
| Approach: | They propose a framework to divide complex instructions into single constraints and prepare appropriate tools to verify responses. |
| Outcome: | The proposed framework doubles Llama3.1-8B’s constraint adherence and triples Mistral-7B’ s performance. |
Stepwise Perplexity-Guided Refinement for Efficient Chain-of-Thought Reasoning in Large Language Models (2025.findings-acl)
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
| 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. |
GAMIC: Graph-Aligned Molecular In-context Learning for Molecule Analysis via LLMs (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Current methods for retrieving large language models rely on molecule feature similarity, such as Morgan fingerprints, which do not adequately capture the global molecular and atom-binding relationships. |
| Approach: | They propose a self-supervised learning technique that embeds demonstration examples into the input prompt. |
| Outcome: | The proposed technique outperforms simple Morgan-based retrieval methods across tasks by up to 45%. |
Image Corruption-Inspired Membership Inference Attacks against Large Vision-Language Models (2026.eacl-long)
Copied to clipboard
| Challenge: | Large vision-language models (LVLMs) are trained on large-scale datasets, which can pose privacy risks if training images contain sensitive information. |
| Approach: | They propose to detect whether a target image is used to train LVLMs by using image-text pairs and single-modality content to detect image-related data. |
| Outcome: | The proposed methods detect whether a target image is used to train the LVLM on large-scale datasets. |
Decoding Time Series with LLMs: A Multi-Agent Framework for Cross-Domain Annotation (2026.findings-eacl)
Copied to clipboard
Minhua Lin, Zhengzhang Chen, Yanchi Liu, Xujiang Zhao, Zongyu Wu, Junxiang Wang, Xiang Zhang, Suhang Wang, Haifeng Chen
| Challenge: | Time series data is ubiquitous across various domains, including manufacturing, finance, and healthcare. |
| Approach: | They propose a multi-agent system to generate general and domain-specific annotations for time series data. |
| Outcome: | The proposed system outperforms existing methods on synthetic and real-world datasets. |
A General Framework to Enhance Fine-tuning-based LLM Unlearning (2025.findings-acl)
Copied to clipboard
Jie Ren, Zhenwei Dai, Xianfeng Tang, Hui Liu, Jingying Zeng, Zhen Li, Rahul Goutam, Suhang Wang, Yue Xing, Qi He, Hui Liu
| Challenge: | Existing approaches to remove copyrighted and privacy-sensitive data from Large Language Models (LLMs) have been proposed to remove specific data from LLMs without requiring full retraining. |
| Approach: | They propose a general framework that enhances the utility of fine-tuning-based methods by distinguishing target data and suppressing related generations. |
| Outcome: | The proposed framework improves the unlearning and utility of fine-tuning-based methods by distinguishing the target data and suppressing related generations. |
InfuserKI: Enhancing Large Language Models with Knowledge Graphs via Infuser-Guided Knowledge Integration (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models have exceptional capabilities in open generation, yet they encounter difficulties with tasks that require intensive knowledge. |
| Approach: | They propose a framework that integrates unknown knowledge into LLMs without overlap . they propose integrating domain-specific knowledge graphs into Llms to reduce knowledge forgetting . |
| Outcome: | The proposed framework outperforms state-of-the-art baselines in integrating new knowledge into LLMs. |
Learning with Less: Knowledge Distillation from Large Language Models via Unlabeled Data (2025.findings-naacl)
Copied to clipboard
Juanhui Li, Sreyashi Nag, Hui Liu, Xianfeng Tang, Sheikh Muhammad Sarwar, Limeng Cui, Hansu Gu, Suhang Wang, Qi He, Jiliang Tang
| Challenge: | Large Language Models (LLMs) have demonstrated superior language understanding abilities in many real-world NLP applications. |
| Approach: | They propose a learning-based sample selection method that incorporates signals from both teacher and student to enhance model performance. |
| Outcome: | The proposed method improves model performance across datasets with higher data efficiency. |
SUA: Stealthy Multimodal Large Language Model Unlearning Attack (2025.emnlp-main)
Copied to clipboard
| Challenge: | Multimodal Large Language Models (MLLMs) trained on massive data may memorize sensitive personal information and photos, posing privacy and copyright concerns. |
| Approach: | They propose a framework that learns a universal noise pattern to recover unlearned information from MLLMs. |
| Outcome: | The proposed framework learns a universal noise pattern and can reveal unlearned content when applied to images. |
A Functionality-Grounded Benchmark for Evaluating Web Agents in E-commerce Domains (2026.acl-long)
Copied to clipboard
| 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. |
Exposing Privacy Risks in Graph Retrieval-Augmented Generation (2026.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have limitations such as generating factually incorrect information (hallucinations) Retrieval-Augmented Generation (RAG) is a powerful paradigm for enhancing LLMs with external, up-to-date knowledge. |
| Approach: | They investigate the data extraction vulnerabilities of Graph RAG systems by executing tailored attacks on them. |
| Outcome: | The proposed attacks exploit the vulnerability of Graph RAG systems to leak raw text and structured data. |
Graph Chain-of-Thought: Augmenting Large Language Models by Reasoning on Graphs (2024.findings-acl)
Copied to clipboard
Bowen Jin, Chulin Xie, Jiawei Zhang, Kashob Kumar Roy, Yu Zhang, Zheng Li, Ruirui Li, Xianfeng Tang, Suhang Wang, Yu Meng, Jiawei Han
| Challenge: | Existing studies suggest augmenting LLMs with external text corpora to alleviate hallucination problems. |
| Approach: | They propose to augment large language models with text units retrieved from external knowledge corpora to alleviate the issue. |
| Outcome: | The proposed framework outperforms baselines on GRBench with three LLMs and shows that iterative reasoning outperformed the baselines. |
Do Multimodal RAG Systems Leak Data? A Comprehensive Evaluation of Membership Inference and Image Caption Retrieval Attacks (2026.findings-acl)
Copied to clipboard
| Challenge: | Multimodal retrieval-augmented generation (mRAG) pipelines are becoming more popular for vision-centric tasks. |
| Approach: | They propose to use a visual asset as a trigger to leak data from a model prompt. |
| Outcome: | The proposed pipelines can connect private datasets and improve model performance, but they can leak private information from them. |
Universal Prompt Optimizer for Safe Text-to-Image Generation (2024.naacl-long)
Copied to clipboard
| Challenge: | Existing studies based on image checker, model fine-tuning and embedding blocking are impractical in real-world applications. |
| Approach: | They propose a novel reward function measuring toxicity and text alignment of generated images and train the optimizer through Proximal Policy Optimization. |
| Outcome: | The proposed model reduces the likelihood of various models in generating inappropriate images, with no significant impact on text alignment. |
A Reward-Guided Dual-Phase Framework for Adaptive Inference-Time Reasoning (2026.findings-acl)
Copied to clipboard
Yingqian Cui, Zhenwei Dai, Pengfei He, Bing He, Hui Liu, Zhan Shi, Xianfeng Tang, Jingying Zeng, Suhang Wang, Yue Xing, Jiliang Tang, Benoit Dumoulin
| Challenge: | Large Language Models (LLMs) have made strong progress in reasoning. |
| Approach: | They propose a dual-phase test-time scaling framework that separates planning and execution and performs search over each phase independently. |
| Outcome: | Experiments on math reasoning and code generation benchmarks show that the proposed approach improves accuracy while reducing redundant computation. |
Graph-Assisted Large Language Models: A Perspective on Mitigating Intrinsic Limitations (2026.findings-acl)
Copied to clipboard
Haitong Luo, Fali Wang, Weiyao Zhang, Xianren Zhang, Zhiwei Zhang, Tianxiang Zhao, Minhua Lin, Jiahao Zhang, Hui Liu, Xianfeng Tang, Qi He, Suhang Wang, Xuying Meng, Yujun Zhang
| Challenge: | Large language models exhibit intrinsic limitations such as knowledge cutoff, single-threaded reasoning that hinders finer-grained branch and aggregation, and rigid collaboration mechanisms that struggle to coordinate specialized capabilities. |
| Approach: | They propose a taxonomy spanning *Graph-Assisted Knowledge Augmentation*, *Graph Assisted Reasoning and Planning*, and *Graphed LLM Collaboration*. |
| Outcome: | The proposed models show that graphs can augment and correct LLMs and support dynamic coordination among experts and agents in collaborative settings. |