開催場所：講堂多目的ルームI および Web（※学外の方も参加可能です）
講演者：Thomas Trappenberg先生（Dalhousie University）
題目： Machine learning, causal reasoning, and human cognition: Where to go from here
Large language models (LLMs) will have a strong impact on our society, and there is an urgent need for better preparation. This is particularly true in education, including how to use these new tools and how to prepare our society for them. While there are huge opportunities, we also face new challenges. For example, the conversational style of chatbots seems often mistaken as causal reasoning. Researchers in machine learning and causal reasoning need to address such misconceptions and clarify transformer abilities. This includes contrasting LLMs with brain processes and the nature of human cognition. I will ask how we can build causal models in the brain that underlie model-based decision making. These would be important steps to understand human cognition and should enable developments of explainable AI.
Dr. Thomas Trappenberg is professor of computer science at Dalhousie University in Halifax, Canada. With a PhD in theoretical Physics from Aachen University, Germany, he has been a postdoctoral fellow at the RIKEN Brain Science Institute and at Oxford University. He is also the Director of AI at Alentic Microscience Inc, a company that develops a lensless microscope for portable blood analysis. Thomas is the author of the text books Fundamental of computational Neuroscience and Fundamentals of Machine Learning, both published at Oxford University Press.