Verifiable and safe multi-agent reinforcement learning

Funder	Engineering and Physical Sciences Research Council
Recipient Organization	Imperial College London
Country	United Kingdom
Start Date	Jan 01, 2025
End Date	Jun 29, 2028
Duration	1,275 days
Number of Grantees	1
Roles	Supervisor
Data Source	UKRI Gateway to Research
Grant ID	`2929173`

Grant Description

The research proposed here aims to enhance and further investigate safe reinforcement learning in the multi-agent context. To deliver on the objectives in question, it will require a cohesive framework that is robust to various forms of multi-agent reinforcement learning as well as being end-to-end (functioning throughout training and deployment).

safe and trusted AI