Speaker: Chao Xu (Associate Professor, School of Philosophy, Shanxi University)

Moderator: Liying Zhang (Professor, Institute of Philosophy, Chinese Academy of Sciences)

Time: Sunday, January 14, 2024, 14:30 - 16:30

Venue: Lecture Hall, Institute of Philosophy, Chinese Academy of Sciences (4th Floor, South Building, Building 4, Software Park, No. 4 Zhongguancun South Street)

Organizer: Institute of Philosophy, Chinese Academy of Sciences (CASIP)

Abstract:
In his 1988 paper, Hector Levesque posed the question, "Can a crocodile jump hurdles?" to examine machines' capacity for commonsense reasoning. Addressing such questions has long been considered foundational for enabling machines to simulate more complex human behaviors. For decades, the academic community has faced significant challenges in devising methods for machines to effectively answer such questions.

With the development of large language models like ChatGPT, machines are now able to understand natural language effectively and learn commonsense knowledge and reasoning patterns from vast amounts of text data, enabling them to tackle such questions with greater success.

This lecture will begin with the problem of commonsense reasoning, followed by an exploration of the symbolic and connectionist approaches to commonsense reasoning, analyzing the difficulties and challenges faced by both. Finally, a comparative analysis of these two approaches will be presented.

Speaker Bio:
Chao Xu holds a Ph.D. in Philosophy from Peking University and is an Associate Professor at the School of Philosophy, Shanxi University. His main research interests include logic and artificial intelligence, logic and cognition, philosophical logic, and the philosophy of artificial intelligence.

Powered by Froala Editor