Formal Methods for Artificial Agents

Speaker:  Sasha Rubin - Università di Napoli
  Monday, June 11, 2018 at 4:30 PM - Aula G
Systems built on the insights of Artificial Intelligence are increasingly
deployed in the world as agents, e.g., software agents negotiating on our
behalf on the internet, robots exploring new and dangerous environments, bots
playing games with humans. There is an obvious need for humans to be able to
control and predict the behaviour of such systems.

This is made all the more complicated since: 1. agents are often deployed with other agents leading to multi-agent systems, 2. agent behaviour extends into the future, leading to the need for reasoning about time, 3. agents may have uncertainty about the state, or even the structure, of other agents and the environment, leading to the need to reason about knowledge, 4. agents are often “self-interested”, leading to the need to reason about strategies.

I will summarise recent theoretical advances, based on logic and automata
theory, that shed light on the verification and synthesis of systems of
artificial agents.

Contact person: Massimo Merro

Contact person

Publication date
May 23, 2018