Training Generalizable Collaborative Agents Via Strategic Risk Aversion

Abstract

Many emerging agentic paradigms require agents to collaborate with one another (or people) to achieve shared goals. Unfortunately, existing approaches to learning policies for such collaborative problems produce brittle solutions that fail when paired with new partners. We attribute these failures to a combination of free-riding during training and a lack of strategic robustness. To address these problems, we study the concept of strategic risk aversion and interpret it as a principled inductive bias for generalizable cooperation with unseen partners. While strategically risk-averse players are robust to deviations in their partner's behavior by design, we show that, in collaborative games, they also (1) can have better equilibrium outcomes than those at classical game-theoretic concepts like Nash, and (2) exhibit less or no free-riding. Inspired by these insights, we develop a multi-agent reinforcement learning (MARL) algorithm that integrates strategic risk aversion into standard pol

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyqu2026training

Related papers

• Learning Generalizable Risk-sensitive Policies To Coordinate In Decentralized Multi-agent General-sum Games (2022)0.00
• Cautiously-optimistic Knowledge Sharing For Cooperative Multi-agent Reinforcement Learning (2023)5.84
• Risk-sensitive Multi-agent Reinforcement Learning In Network Aggregative Markov Games (2024)0.00
• The Price Of Paranoia: Robust Risk-sensitive Cooperation In Non-stationary Multi-agent Reinforcement Learning (2026)0.00
• Loss Aversion Fosters Coordination Among Independent Reinforcement Learners (2019)0.00
• Collaborating With Humans Without Human Data (2021)0.00
• Toward Risk-based Optimistic Exploration For Cooperative Multi-agent Reinforcement Learning (2023)0.00
• Optimism As Risk-seeking In Multi-agent Reinforcement Learning (2025)0.00