Satchit Chatterji
I’m a PhD student at the University of Amsterdam, supervised by dr. Erman Acar (UvA), dr. Shihan Wang (UU) and dr. Giovanni Sileno (UvA). My current research lies in the intersection of logic, reinforcement learning and game theory, aiming to make human-AI systems that are guaranteed to be safe and cooperative. Prior to this I did my Master’s in AI at the UvA and my Bachelor’s in AI at the University of Groningen, both focusing on neurosymbolic methods. My eventual objective is to try and bridge the gap between the strong structure of logic and the exceptional learning power of neural networks, and to apply these methods to enable safe and human-understandable AI. If you’ve read this far, email me your favoriteboard game, esoteric musician/artist, a picture of your pet, or an interesting idea you’d like to collaborate with me on!
Roles in Hybrid Intelligence: PhD Student
Challenge: Collaboration & Synergy / Trust & Trustworthiness
Hybrid Intelligence Project(s):
- 3.9 Neurosymbolic AI for Safety and Norms in Multiagent Reinforcement Learning
Research interests: Reinforcement Learning, Neurosymbolic AI, Game Theory, Multi-Agent Systems, AI Safety
Affiliations:
- Socially Intelligent Artificial Systems Group, Institute of Informatics, University of Amsterdam
Other information: Google Scholar profile
Recent paper:
Chatterji, S., & Acar, E. (2025). Think smart, act SMARL! Analyzing Probabilistic Logic Driven Safety in Multi-Agent Reinforcement Learning.To appear in Proceedings of the 28th European Conference on Artificial Intelligence (ECAI 2025). Preprint available at https://arxiv.org/abs/2411.04867

