Large Language Models meet Logical Reasoning
The 40th Annual AAAI Conference on Artificial Intelligence (AAAI) Tutorial
2:00pm-6:00pm, January 20, 2026, Singapore EXPO
Large Language Models meet Logical Reasoning
The 40th Annual AAAI Conference on Artificial Intelligence (AAAI) Tutorial
2:00pm-6:00pm, January 20, 2026, Singapore EXPO
As Geoffrey Hinton recently remarked, the essence of intelligence is reasoning, which is done by using symbolic rules to manipulate symbolic expressions, and learning can wait, understanding how knowledge is represented in symbolic expressions must come first. This half-day tutorial at AAAI’26 focuses on how to benefit LLMs via symbolic expressions and reasoning. This would improve LLM reasoning capabilities on complex logical problem requiring sophisticated deductive, inductive, or abductive reasoning, also avoid to produce responses that contradict themselves across multiple relevant questions. For symbolic expressions, we introduce effective prompts that translate natural language into different symbolic languages. For symbolic reasoning, we cover external solver-based methods, prompt-based methods, and fine-tuning methods, in which solver-based methods has strong logical reasoning capabilities but sensitive to translation error, while prompt-based methods are opposite. Lastly, we discuss promising research directions, such as which symbolic language should be used and how to develop reasoning method robust to translation error.
Survey Paper:
Fengxiang Cheng, Haoxuan Li, Fenrong Liu, Robert van Rooij, Kun Zhang, Zhouchen Lin. Empowering LLMs with Logical Reasoning: A Comprehensive Survey. In IJCAI, 2025.
Paper Link: https://arxiv.org/pdf/2502.15652
Citation:
@inproceedings{cheng2025empowering,
author={Cheng, Fengxiang and Li, Haoxuan and Liu, Fenrong and Van Rooij, Robert and Zhang, Kun and Lin, Zhouchen},
title={Empowering LLMs with Logical Reasoning: A Comprehensive Survey},
booktitle={Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence},
year={2025}
}
Time: 2:00pm-6:00pm, January 20, 2026
Location: Singapore EXPO, Singapore
Fenrong Liu is a Changjiang Distinguished Professor at Tsinghua University, the Amsterdam-China Logic Visiting Chair at the University of Amsterdam, and Co-Director of the Tsinghua-UvA Joint Research Centre for Logic. She has been elected a Member of the Institut International de Philosophie (IIP) and a Corresponding Member of the Académie Internationale de Philosophie des Sciences (AIPS). Her research interests include logical reasoning for large language models (LLMs), neuro-symbolic AI, multi-agent systems, and preference logic. She has held visiting positions at Harvard University and Keio University, and was a Berggruen Fellow at Stanford University, and served as General Chair of PRICAI 2021 and Program Chair of PRICAI 2023. She is a steering committee member of TARK, PRICAI, LAMAS, and AWPL, and serves on the editorial boards of Synthese, Studia Logica, Global Philosophy, Theoria, and Topoi.
Haoxuan Li is an Assistant Researcher at Peking University and an affiliated Research Fellow at Tsinghua-UvA Joint Research Center for Logic. He pursued his PhD majoring in data science at Peking University from the age of 19. His research interests include causal inference, logical reasoning of LLMs, and recommender system. He has more than 60 publications appeared in top-tier conferences including ICML, NeurIPS, ICLR, AAAI, IJCAI, CVPR, ICDE, ACM KDD, ACM Web Conference, ACM SIGIR, ACM Multimedia and EMNLP, reported by MIT Technology Review. He is supported by the Young Scientists Fund of the National Natural Science Foundation of China and Young Elite Scientists Sponsorship Program by China Association for Science and Technology.
Fengxiang Cheng is a Ph.D. student at the University of Amsterdam. She received her master degree from Tsinghua University. Her research interests include LLMs, logical reasoning, and neuro-symbolic AI. Her first-author researches have been published in IJCAI, AAAI and EMNLP, also reported by MIT Technology Review. Moreover, she was selected as a member of IJCAI 2025 Doctoral Consortium (only 10 Ph.D. students).
Chuan Zhou is a Ph.D student at the University of Melbourne and a Research Engineer at MBZUAI. He received his bachelor and master degree from Peking University. His research interests include large language models, causal inference and neuro-symbolic reasoning. He has published several first-author papers in top conferences such as NeurIPS, ACL, and SIGKDD, also served as the PC member in ICML, AAAI, ACM MM, and SIGKDD.
Mingming Gong is a Senior Lecturer at the University of Melbourne and an affiliated Associate Professor with MBZUAI. He has authored and co-authored 100+ research papers on top venues, a best paper finalist in CVPR 2019, and a first prize in the single image depth estimation competition in CVPR 2018. He has also co-organized several workshops/competitions on SDM 2020, ACML 2021, IJCAI 2021, and AAAI 2025. He received the Discovery Early Career Researcher Award from Australian Research Council in 2020 and was selected as an Early Career Spotlight presenter at IJCAI 2020. He has served as area chair (senior PC) for ICML, NeurIPS, ICLR, AAAI, IJCAI, and ICDM.
Zhouchen Lin is a Distinguished Professor in the School of Intelligence Science and Technology at Peking University. He is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE), Fellow of the International Association of Pattern Recognition (IAPR), Fellow of the Asia-Pacific Artificial Intelligence Association (AAIA), ICML Board Member, ACM Distinguished Member and AAAI Senior Member. He earns the Okawa Research Grant and Microsoft SPOT Award. He is a senior area chair (or senior program committee member) of ICML, NeurIPS, ICLR, AAAI and IJCAI, also an associate editor-in-chief of the IEEE Transactions on Pattern Analysis and Machine Intelligence and an associate editor of the International Journal of Computer Vision.
BibTeX
@article{aaai26-logicllm-tutorial,
author={Liu, Fenrong and Li, Haoxuan and Cheng, Fengxiang and Zhou, Chuan and Gong, Mingming and Lin, Zhouchen},
title={AAAI 2026 Tutorial:Large Language Models meet Logical Reasoning},
journal={AAAI 2026},
year={2026}
}