Large Language Models (LLMs) have emerged as powerful tools for generating content and facilitating information seeking across diverse domains. While their integration into conversational systems opens new avenues for interactive information-seeking experiences, their effectiveness is constrained by their knowledge boundaries—the limits of what they know and their ability to provide reliable, truthful, and contextually appropriate information. Understanding these boundaries is essential for maximizing the utility of LLMs for real-time information seeking while ensuring their reliability and trustworthiness.
In this tutorial, we will explore the taxonomy of knowledge boundary in LLMs, addressing their handling of uncertainty, response calibration, and mitigation of unintended behaviors that can arise during interaction with users. We will also present advanced techniques for optimizing LLM behavior in generative information-seeking tasks, ensuring that models align with user expectations of accuracy and transparency. Attendees will gain insights into research trends and practical methods for enhancing the reliability and utility of LLMs for trustworthy information access.
Our tutorial was held on July 13 (all the times are based on CET).
| Time | Section | Slides |
|---|---|---|
| 09:00—09:30 | Section 1: Introduction + Taxonomy | [Slides] |
| 09:30—10:00 | Section 2: Undesired Behaviors of LLMs | [Slides] |
| 10:00—10:30 | Section 3: Identification of Knowledge Boundary | [Slides] |
| 10:30—11:00 | Coffee Break | - |
| 11:00—11:20 | Section 4: Mitigation of Out-of-Boundary Knowledge: Outward Boundary | [Slides] |
| 11:20—11:50 | Section 5: Mitigation of Out-of-Boundary Knowledge: Parametric Boundary | [Slides] |
| 11:50—12:10 | Section 6: Mitigation of Out-of-Boundary Knowledge: Universal Boundary | [Slides] |
| 12:10—12:30 | Section 7: Open Challenges and Beyond + Q&A | [Slides] |
Yang Deng is an Assistant Professor at Singapore Management University. His research lies in information retrieval (IR) and natural language processing (NLP), especially for trust and reliability in LLMs. He has published over 60 papers on relevant topics at top venues such as SIGIR, WWW, ACL, EMNLP, ICLR, TOIS, TKDE, and serves as Area Chairs. He received the Google Southeast Asia Research Awards in 2024 for his excellent research on trustworthy AI. He has rich experience in organizing tutorials at top conferences, including SIGIR 2024, WWW 2024, and ACL 2023.
Moxin Li is a final-year PhD candidate at National University of Singapore. Her research focuses on IR and NLP, especially for LLM trust and evaluation. She has published over 10 papers at top conferences including SIGIR, WWW, ACL, EMNLP, etc.
Liang Pang is an Associate Professor at the Institute of Computing Technology, Chinese Academy of Sciences. He is renowned for his expertise in trustworthy LLMs and text matching in IR. He has published about 60 papers in top journals and conferences, including SIGIR, WWW, ACL, EMNLP, etc, and received the Best Paper Runner-up of CIKM 2017, the Best Paper Honorable Mention of SIGIR 2024. He has delivered multiple tutorials at SIGIR 2021, WSDM 2021, and KDD 2024.
Wenxuan Zhang is an Assistant Professor at Singapore University of Technology and Design. Before this, he was a research scientist at Alibaba DAMO Academy, Singapore. His primary research areas are NLP and trustworthy AI, with a special aim to advance inclusive NLP, supporting diverse languages and cultures. He has published over 40 papers in top-tier conferences and journals, including SIGIR, WWW, ICLR, NeurIPS, ACL, EMNLP. He also regularly serves on the (senior) program committees of multiple leading conferences and journals. He organized a tutorial at IJCAI 2023.
Wai Lam is a Professor at the Department of Systems Engineering and Engineering Management, The Chinese University of Hong Kong, where he is also the Division Head and Department Chairman. His research interests include intelligent information retrieval and text mining. He has authored or co-authored more than 100 papers in premier conferences and journals (SIGIR, WWW, ACL, EMNLP, etc). He regularly serves as the Senior PC or Area Chair of these conferences. He receives multiple paper awards, including ACL 2021 Outstanding Paper Awards, EACL 2023 Outstanding Paper Awards, and ACL 2024 Area Chair Awards.
@inproceedings{sigir25-knowledge-boundary-tutorial,
title={Unveiling Knowledge Boundary of Large Language Models for Trustworthy Information Access},
author={Deng, Yang and Li, Moxin and Liang, Pang and Zhang, Wenxuan and Lam, Wai},
booktitle={Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '25), July 13--18, 2025, Padua, Italy},
year={2025}
}