University of Birmingham > Talks@bham > Artificial Intelligence and Natural Computation seminars > Multiagent Planning, Learning & Influences

Multiagent Planning, Learning & Influences

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Lars Kunze.

Speaker’s homepage:

Host: Dr. Bruno Lacerda

ABSTRACT : Multiagent systems hold great promise as they potentially offer more efficient, robust and reliable solutions to a great variety of real-world problems, such as the control of networks of traffic lights or pumping stations, decision making in teams of robots, etc. Consequently, multiagent planning and learning have been important research topics in artificial intelligence for nearly two decades.

In particular, this talk will focus on decision-theoretic approaches to multiagent planning and learning, formulated as (special cases of) decentralized partially observable Markov decision processes (Dec-POMDPs). These methods have the benefit of having clearly quantified objectives, thus enabling principled ways to trade off, e.g., information gathering versus exploitation of that information. Despite great progress, the scalability of these approaches remains a core issue.

In this talk I will give an overview of some of the approaches to improve the scalability of multiagent learning and planning that I have pursued in recent years. A common thread in these is the attempt to capture the locality in the way that agents may influence one another. Formalizations of such ‘influences’ have lead to vast improvements in planning tractability. For instance, I will talk about a recent approach that, for the first time ever, enabled us to verify the (near-)optimality of heuristic solutions of Dec-POMDPs with 100s of agents. I will argue that influences will be critical to the advancement of multiagent learning too.

This talk is part of the Artificial Intelligence and Natural Computation seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


Talks@bham, University of Birmingham. Contact Us | Help and Documentation | Privacy and Publicity.
talks@bham is based on from the University of Cambridge.