Title: Formal Methods in AI and Machine Learning

Description: Speaker: Dr. Blaise Genest

Time: Friday, 21 February 2020, 11:30am
Venue: Department of Computer Science and Engineering, Room No. 109, 01st Floor, New CSE/CC Building

Abstract:
In this talk, we describe several ways to foster trust in AI tools. Indeed, the most prominent current AI tools (Neural Networks, Reinforcement learning) are in the same time very efficient, but also very complex to be understood by humans. And these tools fail sometimes, with potentially catastrophic outcomes. Recently, the new paradigm of explainable AI has emerged. While explainable AI is necessary, it is also not sufficient for complete trust: what if we are not convinced by the explanations? Or what if we are wrongly convinced by it? Also, humans cannot check every split-second decisions taken by AI tools. In order to complement explainable AI, we advocate for advances in Formal Methods and AI to provide guarantees about their behaviors. In this talk, we will particularly focus on guarantees we can bring when Learning a stochastic system. More precisely, we study several simple algorithm to learn Discrete Time Markov Chains (frequency estimation and Laplace smoothing) and provide statistical PAC guarantees over global behavior (reachability probabilities…) of the learnt model wrt the original system.

Speaker Profile:
Blaise Genest is a former student of ENS Cachan/Paris-Saclay (Mathematics and Computer Science, 1999-2003). He received his Ph.D. in Computer Science from University Paris 7, France (2004). He is a CNRS senior researcher with IRISA, Rennes, France. He spent 3 years in Singapore in 2010-2012, working with NUS. He is working at the interface of Formal Methods and AI. He organized the 2nd International Workshop on FM and AI in Rennes in 2019, which attracted around 50 researchers from 11 countries. The 3rd workshop FMAI’2020 will take place at Imperial College.

Organization:
IRISA, Rennes



Host:
Prof. Akshay S.

Date: Friday, 21 February, 2020
Time: 11:30am IST
Access: Public
Category: Talk*
Created by: Department Calendar
Updated: Tuesday, 30 November, 2021 8:44am IST
Send Reminder: Yes  -  46 hours 6 minutes before start
Participants: Department Calendar
<office@cse.iitb.ac.in> (External User)
_NUC_department <all@cse.iitb.ac.in> (External User)