Re-Acting to Partner's Actions with the Social Intrinsic Motivation of Transfer Empowerment

Tessa van der Heiden, Herke van Hoof, Efstratios Gavves, Christoph Salge

Research output: Chapter in Book/Report/Conference proceedingConference contribution


We consider multi-agent reinforcement learning (MARL) for cooperative communication and coordination tasks. MARL agents can be brittle because they can overfit their training partners’ policies. This overfitting can produce agents that adopt policies that act under the expectation that other agents will act in a certain way rather than react to their actions. Our objective is to bias the learning process towards finding reactive strategies towards other agents’ behaviors. Our method, transfer empowerment, measures the potential influence between agents’ actions. Results from three simulated cooperation scenarios support our hypothesis that transfer empowerment improves MARL performance. We discuss how transfer empowerment could be a useful principle to guide multi-agent coordination by ensuring reactiveness to one’s partner.
Original languageEnglish
Title of host publicationALIFE 2022: The 2022 Conference on Artificial Life
PublisherMIT Press
Number of pages9
Publication statusAccepted/In press - 28 Apr 2022
EventArtificial Life 2022 - online , Trento, Italy
Duration: 18 Jul 202222 Jul 2022


ConferenceArtificial Life 2022
Abbreviated titleAlife 2022
Internet address


Dive into the research topics of 'Re-Acting to Partner's Actions with the Social Intrinsic Motivation of Transfer Empowerment'. Together they form a unique fingerprint.

Cite this