Bálint Gyevnár
PhD student in AI safety and explainable AI

Hi, I am Bálint. Thanks for checking on my home page! (Bluesky)
(My name is pronounced BAH-lint [baːlint])
My primary research area is explainable multi-agent reinforcement learning. I like to describe this as the study of giving interacting AI agents the ability to explain themselves.
I am primarily interested in how we can explain complex emergent behaviour in multi-agent systems (MAS) via the use of counterfactual reasoning, and how this in turn can be used to calibrate trust and verify the safety of MAS.
I also work on bridging the epistemic foundations and research problems of AI ethics and safety to foster cross-disciplinary collaboration.
I am a member of the Autonomous Agents Research Group, supervised by Shay Cohen and Chris Lucas. I was previously supervised by Stefano Albrecht.
news
Jun 18, 2025 | I have attended the 2025 Bridging Responsible AI Divides (BRAID) Gathering in Manchester. |
---|---|
Jun 11, 2025 | I have attended RLDM 2025, the Multi-disciplinary Conference on Reinforcement Learning and Decision Making, in Dublin, Ireland, where I have presented a poster on our objective metrics for explainable RL paper. |
Jun 07, 2025 | I gave a talk and presented a poster at the 9th CHAI Workshop on “AI Safety for Everyone”. |
May 26, 2025 | New preprint paper titled: Integrating Counterfactual Simulations with Language Models for Explaining Multi-Agent Behaviour. |
Mar 21, 2025 | I am co-organising a workshop on “Evaluating Explainable AI and Complex Decision-Making” co-located with ECAI ‘25. Call for papers found here. |
Feb 25, 2025 | New journal paper at Nature Machine Intelligence: AI Safety for Everyone. |
Feb 10, 2025 | I attended IASEAI ‘25: the inaugural conference of the International Association for Safe and Ethical AI. Program available here. |