Research map / Reinforcement learning / Visual and representation RL
Adaptable Agent Populations via a Generative Model of Policies
Kenneth Derek, Phillip Isola
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add action-distribution diversity regularization to PPO | ADAP on Farmworld niche specialization · Farmworld niche specialization | episode reward | Improved |
| Add action-distribution diversity regularization to PPO | ADAP on Farmworld niche specialization · Farmworld niche specialization | per-agent specialization | Improved |
| Add diversity regularization | PPO on Farmworld environmental ablations · Farmworld environmental ablations | reward | Improved |
| Add diversity regularization | PPO on Farmworld Speed ablation · Farmworld Speed | reward | No clear effect |
| Add diversity regularization | PPO against Markov Soccer opponents · Markov Soccer | Wins - Losses | Improved |
| Use a multiplicative latent-integration architecture instead of concatenation (ablation) | ADAP on Farmworld niche specialization · Farmworld niche specialization | policy-space multi-modality | Improved |
| Use a concatenation architecture instead of multiplicative latent integration (ablation) | ADAP on Farmworld niche specialization · Farmworld niche specialization | niche learning consistency | Hurt |
| Remove diversity regularization (ablation) | PPO on Farmworld environmental ablations · Farmworld environmental ablations | reward | Hurt |
| Use a shared value and policy network instead of separate networks (ablation) | ADAP · Farmworld and other experiments | policy-space diversity | Hurt |
Closest papers by meaning
- Enabling Adaptive Agent Training in Open-Ended Simulators by Targeting Diversity (NeurIPS 2024 · similarity 0.88)
- Efficient Adaptation in Mixed-Motive Environments via Hierarchical Opponent Modeling and Planning (ICML 2024 · similarity 0.87)
- GenPlan: Generative Sequence Models as Adaptive Planners (AAAI 2025 · similarity 0.87)
- Learning Diverse Risk Preferences in Population-Based Self-Play (AAAI 2024 · similarity 0.87)
- Agent World Model: Infinity Synthetic Environments for Agentic Reinforcement Learning (ICML 2026 · similarity 0.87)
- Revisiting Multi-Agent World Modeling from a Diffusion-Inspired Perspective (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
