Social preferences enable cooperation without seeing others’ rewards
Self-Referenced Social Preferences: Cooperation without Observing Others Rewards
Artificial IntelligenceMultiagent Systems
Summary
Sometimes, people and computers work together best when they care about each other’s success. This paper shows that computer agents can learn to cooperate by guessing how others feel about their own success, even without knowing their exact rewards. The researchers created a method where agents use their own sense of reward to interpret how well others are doing based on what they see. These techniques helped agents cooperate in tricky group situations better than if they saw others’ rewards directly.
What this means in practice
- •For multi-agent system developers: Design cooperation strategies for agents without needing direct access to others’ private rewards, enabling more privacy-respecting multi-agent environments.
- •For robotics teams: Enable robots to learn cooperative behaviors by observing others’ actions and outcomes rather than requiring shared reward signals, useful in decentralized settings.
Authors
Mohamed Ayman Mohamed, Harshil Kotamreddy, Marcos Menon Jose
Abstract
Social preferences can promote cooperation in multi-agent reinforcement learning, but existing approaches often require agents to observe the rewards of their peers. In many real-world interactions, however, an agent can, as humans do, observe others' behavior and outcomes without access to their private reward signals. We introduce self-referenced social preferences, in which each agent learns a model of its own reward, applies it to other agents' observed transitions to assess their outcomes from its own perspective, and feeds these self-referenced assessments into standard social preferences. We study two ways to incorporate these assessments: modifying the learning reward, or using them to weight policy updates. We evaluate the approach on three sequential social dilemmas, Escape Room, Clean Up, and Commons Harvest, which require volunteering, public-good contribution, and resource restraint, respectively. Across all three environments, agents learn cooperative behavior without observing others' rewards, including in settings where independent learners fail to cooperate, and frequently achieve more equitable divisions of jointly produced returns than agents with access to true rewards. The effective integration point depends on the social preference: inequity aversion works best in the reward together with a value look-ahead, whereas a purely benevolent preference benefits from policy-update weighting. Under partial observability, the policy-update approach continues to support cooperation. These results show that explicit access to other agents' reward signals is not necessary for learning cooperative behavior: social preferences can instead be grounded in self-referenced assessments of others' outcomes derived from their observed behavior.