LOQA: Learning with Opponent Q-Learning Awareness

arXiv:2405.01035v1 Announce Type: cross
Abstract: In various real-world scenarios, interactions among agents often resemble the dynamics of general-sum games, where each agent strives to optimize its own utility. Despite the ubiquitous relevance of such settings, decentralized machine learning algorithms have struggled to find equilibria that maximize individual utility while preserving social welfare. In this paper we introduce Learning with Opponent Q-Learning Awareness (LOQA), a novel, decentralized reinforcement learning algorithm tailored to optimizing an agent's individual utility while fostering cooperation among …

abstract agent agents algorithms arxiv cs.ai cs.gt cs.lg decentralized dynamics equilibria games general interactions machine machine learning machine learning algorithms q-learning resemble social sum type utility welfare while world