Dr. Marcelo Mendoza
Computer Science Department at Pontificia Universidad Católica de Chile, (Chile)
Artificial Reasoning and ChatGPT
Toward a general AI system by combining NLP with the ability to simulate human-like reasoning
Artificial reasoning is a field of artificial intelligence focused on uncovering the reasoning capabilities of inductive learning models. The emergence of large language models (LLMs) has sparked significant interest in this area, primarily due to their surprising emergent reasoning abilities across various problem types. This talk illustrates the capabilities and limitations of artificial reasoning exhibited by models like ChatGPT and suggests challenges that must be addressed to construct a general artificial intelligence system.
Marcelo Mendoza is an associate professor in the Computer Science Department at Pontificia Universidad Católica de Chile. He received a Ph.D. in Computer Science from the Universidad de Chile. He did a postdoc in Yahoo Research. He is a founder and former President of the Chilean Association of Pattern Recognition. He is a principal researcher at the National Center of Artificial Intelligence (CENIA) and associate researcher at the Millennium Institute for Foundational Research on Data (IMFD). His research is focused on developing AI methods to measure and predict events and processes involving humans. His work includes the study of information dissemination, the spatial distribution of humans within cities, community formation, and the analysis of constructive and destructive interactions among individuals. His contributions are significant in advancing the understanding of how AI can address complex societal challenges, particularly those related to the dynamics of human behavior. He has published over 100 peer-reviewed papers in conference proceedings and top-tier journals. In 2021, he received the Seoul Test of Time award for co-authoring the paper Information Credibility on Twitter, a pioneering study in applying AI to the analysis of misinformation.