IJCAI-25: Dual Venues, Keynotes, & Cutting-Edge AI Research Unveiled
The 34th International Joint Conference on Artificial Intelligence (IJCAI-25), a cornerstone event in the global AI research calendar, is set to convene in Montréal, Canada, from August 16 to 22. This year marks a significant expansion, with a concurrent satellite event scheduled in Guangzhou, China, from August 29 to 31, broadening the conference’s international reach and accessibility.
The main Montréal program promises a comprehensive exploration of AI’s cutting edge, featuring a rich agenda of keynote talks, tutorials, workshops, and competitions, alongside traditional oral and poster presentations. Complementing these are four specialized tracks, delving into critical areas such as AI for social good, the intersection of AI and arts, human-centred AI, and AI’s role in enabling essential technologies. The Guangzhou satellite event mirrors this structure, offering its own series of invited talks, workshops, tutorials, technical sessions, posters, and competitions, ensuring a robust scientific exchange across two continents.
A total of twelve distinguished invited talks are slated for the conference, strategically distributed between the two venues. Montréal will host seven of these insightful sessions. Bernhard Schölkopf is scheduled to present on “From ML for science to causal digital twins,” while Cynthia Rudin will discuss “Interpretable machine learning and AI, John McCarthy and I.” Heng Ji will explore “Science-inspired AI,” and Luc De Raedt will delve into “Neurosymbolic AI: combining data and knowledge,” a topic bridging data-driven and knowledge-based approaches. Renowned AI pioneer Yoshua Bengio is also among the speakers, with his title yet to be confirmed, as is Aditya Grover’s. Rina Dechter will share “Graphical models meet heuristic search: a personal journey into automated reasoning.”
The Guangzhou venue will feature five invited speakers, each contributing to diverse facets of AI. Toby Walsh’s talk title is pending confirmation, but Harry Shum will discuss “Exploring the low altitude airspace: from natural resource to economic engine.” Yew Soon Ong will present on “Physically grounded AI for scientific discovery: from prediction to generative design,” highlighting AI’s application in fundamental science. Shing-Tung Yau will address “Advancing artificial intelligence through modern mathematical theories,” underscoring the mathematical underpinnings of AI, and Wen Gao’s talk title is also awaiting announcement.
The conference’s educational component is robust, with tutorials planned for Montréal from August 16 to 18 and in Guangzhou on August 29. Montréal’s offerings include deep dives into “Scaling LLM Training: Efficient Pre-training & Fine-tuning on AI Accelerators” and “Evaluating LLM-based Agents,” alongside “Beyond Text: Advanced Retrieval Augmented Generation for Complex and Multimodal Data.” Other topics span “Neuroevolution of Intelligent Agents,” “AI Meets Algebra,” “Principles of Self-supervised Learning in the Foundation Model Era,” and “Advances in Time-Series Anomaly Detection.” Practical applications are also covered, such as “Computational Pathology Foundation Models” and tutorials on fairness in AI, including “Fairness in Large Language Models” and “Supervised Algorithmic Fairness in Distribution Shifts.” Guangzhou’s tutorials will focus heavily on large language models (LLMs), with sessions on “LLM-based Role-Playing from the Perspective of Hallucinations,” “Empowering LLMs with Logical Reasoning,” and “Large Language Models for Recommendation.” Further Guangzhou tutorials explore “Multimodal Large Language Model for Visually Rich Document Understanding,” “GUI Agents with Foundation Models,” and “Multi-Modal Generative AI in Dynamic and Open Environment.”
Workshops, integral to fostering focused discussions and collaborations, will also run concurrently with the tutorial programs in both locations. Montréal’s extensive workshop schedule includes specialized gatherings on “Knowledge Representation and Reasoning,” “Deepfake Detection, Localization and Interpretability,” and “Multimodal Knowledge and Language Modeling.” Ethical and societal dimensions of AI are addressed in workshops like “AI for Aging Rehabilitation and Intelligent Assisted Living,” “Social Choice and Learning Algorithms,” “Democracy and AI,” and “Artificial Intelligence for Sustainability.” Other notable Montréal workshops delve into “Federated Learning with Generative AI,” “Trust in the new Agent Societies,” “Explainable Artificial Intelligence (XAI),” and applications in healthcare with “Large Language Models and Generative AI for Health Informatics” and “Empowering Women of Colour in AI-Driven Mental Health Research.” Meanwhile, Guangzhou’s workshops will tackle “Large Language Models for Financial Services,” “Generalizing from Limited Resources in the Open World,” and “Mechanism Design in Social Networks.” Practical applications are further explored in “4D Micro-Expression Recognition for Mind Reading” and “Multimodal Fusion Learning for Remote Physiological Signal Sensing,” alongside “Practical Deep Learning” focusing on robust compressed foundation models.
The dual-venue format of IJCAI-25 underscores the increasingly global and multifaceted nature of AI research. By bringing together leading minds and diverse perspectives across two continents, the conference is poised to showcase the latest breakthroughs, address critical challenges, and foster international collaboration in a field that continues to reshape industries and societies worldwide.