LLM & AI Agent

Research on LLM & AI Agent: 

• • For LLM & AI Agent in Time Series and Education domains, check AI for Education, AI for Time Series.

  • The full list of my publications (with arXiv papers) can be found at Google Scholar.

[Trustworthy AI/LLM/Agent]: Safety, Security, Privacy, Robustness, Interpretability, Fairness, Hallucination, Alignment, etc.

Algorithm & Benchmark & Data & Code: 

• • [ICML'26] SafeSeek: Universal Attribution of Safety Circuits in Language Models, ICML 2026. [arXiv]

  • [ICML'26] Uncovering Hidden Triggers: Backdoor Attribution in Language Models, ICML 2026. [arXiv]

• [CVPR'26] AutoDebias: An Automated Framework for Detecting and Mitigating Backdoor Biases in Text-to-Image Models, CVPR 2026. [arXiv]

  • [ACL'26] Backdoor Collapse: Eliminating Unknown Threats via Known Backdoor Aggregation in Language Models, ACL 2025. [arXiv]

  • [ACL'26] HearSay Benchmark: Do Audio LLMs Leak What They Hear? ACL 2026. [arXiv]

  • [KDD'26] Evaluating RAG Robustness to Symbolic Perturbations, KDD 2026. [arXiv]

• [ICLR'26] DiffuGuard: How Intrinsic Safety is Lost and Found in Diffusion Large Language Models, ICLR 2026. [arXiv]

• [ICLR'26] AudioTrust: Benchmarking The Multifaceted Trustworthiness of Audio Large Language Models, ICLR 2026. [arXiv]

• [NeurIPS'25] SAEMark: Steering Personalized Multilingual LLM Watermarks with Sparse Autoencoders, NeurIPS 2025. [arXiv]

• [ACL'25] NetSafe: Exploring the Topological Safety of Multi-agent Networks, ACL 2025. [arXiv]

• [MM'25] The Eye of Sherlock Holmes: Uncovering User Private Attribute Profiling via Vision-Language Model Agentic Framework, ACM MM 2025. [arXiv]

• [MM'25] Debiasing Multimodal Large Language Models via Penalization of Language Priors, ACM MM 2025. [arXiv] [code]

• AgentSafe: Safeguarding Large Language Model-based Multi-agent Systems via Hierarchical Data Management, arXiv 2025. [arXiv]

  • ChronosAudio: A Comprehensive Long-Audio Benchmark for Evaluating Audio-Large Language Models, arXiv 2026. [arXiv]

• MCPShield: A Security Cognition Layer for Adaptive Trust Calibration in Model Context Protocol Agents, arXiv 2026. [arXiv]

Survey & Position & Perspective: 

• [KDD'25] A Survey on Trustworthy LLM Agents: Threats and Countermeasures, KDD 2025. [arXiv]

• [SMC Magazine'25] R4 Trustworthy Human–Artificial Intelligence Symbiosis, IEEE SMC Magazine 2026. [paper]

  • A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment, arXiv 2025. [arXiv]

• LLM-based Agents Suffer from Hallucinations: A Survey of Taxonomy, Methods, and Directions, arXiv 2025. [arXiv]

• Aligning Multimodal LLM with Human Preference: A Survey, arXiv 2025. [arXiv]

• Evaluating LLMs in Finance Requires Explicit Bias Consideration, arXiv 2026. [arXiv]

[Others]: Agent, LLM, MLLM, VLM, RAG, RecSys, BCI, etc.

Algorithm & Benchmark & Data & Code: 

• [ICML'26] See First, Reason Later: Mutual Information-Guided Reinforcement Learning for Vision-Language Models, ICML 2026.

• [ACL'26] ARVEC: Adaptive Reasoning Model with Vision Understanding and Executable Code, ACL 2026. [arXiv]

• [ACL'26] Scaling Law for Multimodal Large Language Model Supervised Fine-Tuning, ACL 2026. 

  • [AAAI'26] Assemble Your Crew: Automatic Multi-agent Communication Topology Design via Autoregressive Graph Generation, AAAI 2026. [arXiv]  (AAAI Oral, Top 5%) 

  • [AAAI'26] SafeSieve: From Heuristics to Experience in Progressive Pruning for LLM-based Multi-Agent Communication, AAAI 2026. [arXiv]

• [TPAMI'26] Beyond LLaVA-HD: Diving into High-Resolution Large Multimodal Models, IEEE TPAMI, 2026. [arXiv]

• [ICLR'26] RankLLM: Weighted Ranking of LLMs by Quantifying Question Difficulty, ICLR 2026. [arXiv]

• [ICLR'25] MME-RealWorld: Could Your Multimodal LLM Challenge High-Resolution Real-World Scenarios that are Difficult for Humans? ICLR 2025. [arXiv]

• [NeurIPS'25] Improving Nonlinear RNN with Closed-loop Control, NeurIPS 2025. [arXiv] (NeurIPS Spotlight, Top 3.5%) 

• [NeurIPS'25] Can LLMs Correct Themselves? A Benchmark of Self-Correction in LLMs, NeurIPS 2025. 

• [NeurIPS'24] AutoSurvey: Large Language Models Can Automatically Write Surveys, NeurIPS 2024 [arXiv] 

• [AAAI'25] UrbanVLP: Multi-Granularity Vision-Language Pretraining for Urban Socioeconomic Indicator Prediction, AAAI 2025 [paper]

• [EMNLP'25] DynamicNER: A Dynamic, Multilingual, and Fine-Grained Dataset for LLM-based Named Entity Recognition, EMNLP 2025. [arXiv]

• [EMNLP'24] RAGLAB: A Modular and Research-Oriented Unified Framework for Retrieval-Augmented Generation, EMNLP 2024 [arXiv] 

• [WWW'24] UrbanCLIP: Learning Text-enhanced Urban Region Profiling with Contrastive Language-Image Pretraining from the Web, WWW ’24 [arXiv]

Survey & Position & Perspective: 

• [ACL'26] Position: Multimodal Large Language Models Can Significantly Advance Scientific Reasoning, ACL 2026. [arXiv]

• [SPM'25] Brain Foundation Models: A Survey on Advancements in Neural Signal Processing and Brain Discovery, IEEE Signal Processing Magazine, 2025. [arXiv] 

• [ACL'25] A Survey of Mathematical Reasoning in the Era of Multimodal Large Language Model: Benchmark, Method & Challenges, ACL 2025. [arXiv]

• [FnTs'25] Recommender Systems Meet Large Language Model Agents: A Survey, Foundations and Trends® in Privacy and Security, 2025. [link] [Preprint]

  • Position: Multimodal large language models can significantly advance scientific reasoning, arXiv 2025. [arXiv]

  • Topological structure learning should be a research priority for LLM-based multi-agent systems, arXiv 2025. [arXiv]

• A Survey on Post-training of Large Language Models, arXiv 2025. [arXiv]

• Complex networks of AI agentic systems: topology, memory, and update dynamics, arXiv 2026 [arXiv]

• Digital Twin AI: Opportunities and Challenges from Large Language Models to World Models, arXiv 2026 [arXiv]

Workshop on AI Agent (@ KDD, WWW, CIKM, AAAI): 

• KDD Workshop on AI Agent for Information Retrieval (KDD'25-Agent4IR, KDD'26-Agent4IR)

• WWW Workshop on AI Agent for Information Retrieval: Generating and Ranking (WWW'25-Agent4IR)

• AAAI Workshop on AI Agent for Information Retrieval: Generating and Ranking (AAAI'25-Agent4IR) 

• CIKM Workshop on AI Agent for Information Retrieval (CIKM'24-Agent4IR)