AASS Seminar - Can AI help understand human errors?
22 november 2024 13:00 Hörsal L3, Långhuset
For more information about the AASS Seminar Series, please contact:
Alessandro Saffiotti
The research centre AASS arranges a seminar with Nicolas Sabouret, University Paris Saclay.
Remote attendance: https://oru-se.zoom.us/j/65096623293
Abstract
Human error analysis consists in trying to understand what went wrong in the case of an incident. Such analysis is made even more difficult when it comes to modern socio-technical systems, in which human operators interact with complex systems. Accidents can occur even with a fully operational system and rational operators. In such situations, it can be very difficult to identify the cause of the error. As an example, we all know how long the inquiry after an airplane crash can be. In this talk, we will show how AI and logic-based reasoning can help investigators in exploring the numerous scenarios that can explain the accident. We show that human biases can be modeled and used to filter the possible explanations. This work, at the junction between AI and cognitive sciences, gives an example of AI use that does not try to replace human beings, but rather serve as a tool for human enhancement.
Bio
Nicolas Sabouret is full Professor at University Paris-Saclay, conducting his research at LISN, the interdisciplinar joint research institute for Digital Science between Université Paris-Saclay, CNRS, Inria and CentraleSupélec. He is also director of the Graduate School of Computer Science. Nicolas Sabouret is an expert in Artificial Intelligence. His research focuses on human behaviour modeling and simulation, combining AI models with theories of human cognition and decision making from social science. His goal is to contribute to the construction of artificial agents that adopt human-like behaviours. His research is applied to various domains: multi-agent simulation of human activity to study and predict electrical consumption in households; development of conversational agents with social competencies for customer relationship, for social training or for health coaching; modeling and simulation of human factors in critical systems.