mlsanityData to 7 Oct 2026

The Good Shepherd: An Oracle Agent for Mechanism Design

arXiv 2202.10135 ↗First posted 21 Feb 2022cs.MA cs.AI cs.LGcited by 10

Not on the map. Why?

Abstract (arXiv)

From social networks to traffic routing, artificial learning agents are playing a central role in modern institutions. We must therefore understand how to leverage these systems to foster outcomes and behaviors that align with our own values and aspirations. While multiagent learning has received considerable attention in recent years, artificial agents have been primarily evaluated when interacting with fixed, non-learning co-players. While this evaluation scheme has merit, it fails to capture the dynamics faced by institutions that must deal with adaptive and continually learning constituents. Here we address this limitation, and construct agents ("mechanisms") that perform well when evaluated over the learning trajectory of their adaptive co-players ("participants"). The algorithm we propose consists of two nested learning loops: an inner loop where participants learn to best respond to fixed mechanisms; and an outer loop where the mechanism agent updates its policy based on experience. We report the performance of our mechanism agents when paired with both artificial learning agents and humans as co-players. Our results show that our mechanisms are able to shepherd the participants strategies towards favorable outcomes, indicating a path for modern institutions to effectively and automatically influence the strategies and behaviors of their constituents.

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Builds on 11 of its 20 references that are in the map. First posted on arXiv 21 Feb 2022.

Builds on

11 of 20 references.

References · 11

  1. 2016Learning to Communicate with Deep Multi-Agent Reinforcement Learningcited by 561
  2. 2017Evolution Strategies as a Scalable Alternative to Reinforcement Learningcited by 555
  3. 2017Value-Decomposition Networks For Cooperative Multi-Agent Learningcited by 490
  4. 2017Learning with Opponent-Learning Awarenesscited by 218
  5. 2021Collaborating with Humans without Human Datacited by 139
  6. 2018The Mechanics of n-Player Differentiable Gamescited by 118
  7. 2018Bayesian Action Decoder for Deep Multi-Agent Reinforcement Learningcited by 67
  8. 2020Learning to Incentivize Other Learning Agentscited by 55
  9. 2019A Generalized Training Approach for Multiagent Learningcited by 53
  10. 2017Maintaining cooperation in complex social dilemmas using deep reinforcement learningcited by 48
  11. 2017Consequentialist conditional cooperation in social dilemmas with imperfect informationcited by 23

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 1; 2024: 3; 2025: 1; 2026: 2.

10 papers.

  1. 2022Influencing Long-Term Behavior in Multiagent Reinforcement Learningcited by 18
  2. 2022Oracles & Followers: Stackelberg Equilibria in Deep Multi-Agent Reinforcement Learningcited by 12
  3. 2023Scaling Opponent Shaping to High Dimensional Gamescited by 12
  4. 2024Multi-agent cooperation through learning-aware policy gradientscited by 4
  5. 2024Best Response Shapingcited by 3
  6. 2022Developing, Evaluating and Scaling Learning Agents in Multi-Agent Environmentscited by 2
  7. 2025Opponent Shaping in LLM Agentscited by 2
  8. 2026Multi-agent cooperation through in-context co-player inferencecited by 1
  9. 2024Using deep reinforcement learning to promote sustainable human behaviour on a common pool resource problemcited by 0
  10. 2026Deep Incentive Design with Differentiable Equilibrium Blockscited by 0