Research map / Reinforcement learning / Policy gradient and actor-critic
CEM-RL: Combining evolutionary and gradient-based methods for policy search
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| combine CEM with TD3 gradients | CEM-TD3 · HalfCheetah-v2, Hopper-v2, Walker2d-v2, Ant-v2 | mean return: 2940 -> 10725; 1055 -> 3613; 928 -> 4711; 487 -> 4251 | Helped |
| combine CEM with TD3 gradients | CEM-TD3 · Swimmer-v2 | mean return: 351 -> 75 | Hurt |
| add the CEM evolutionary mechanism to multi-actor TD3 (ablation) | CEM-TD3 · HalfCheetah-v2, Hopper-v2, Walker2d-v2, Ant-v2 | mean return: 9662 -> 10725; 2056 -> 3613; 3934 -> 4711; 3567 -> 4251 | Helped |
| add the CEM evolutionary mechanism to multi-actor TD3 (ablation) | CEM-TD3 · Swimmer-v2 | mean return: 76 -> 75 | No clear effect |
| replace DDPG with TD3 (ablation) | CEM-RL · Walker2d-v2, Ant-v2 | mean return: 2865 -> 4711; 2170 -> 4251 | Helped |
| replace DDPG with TD3 (ablation) | CEM-RL · Swimmer-v2 | mean return: 268 -> 75 | Hurt |
| add importance mixing (ablation) | CEM-TD3 · HalfCheetah-v2, Hopper-v2, Walker2d-v2, Ant-v2 | mean return: 10725 -> 10601; 3613 -> 3589; 4711 -> 4420; 4251 -> 4235 | Hurt |
| add importance mixing (ablation) | CEM-TD3 · Swimmer-v2 | mean return: 75 -> 117 | Helped |
| add action-space Gaussian noise (ablation) | CEM-TD3 · HalfCheetah-v2, Hopper-v2, Walker2d-v2, Swimmer-v2, Ant-v2 | mean return: 10725 -> 11006; 3613 -> 3541; 4711 -> 4542; 75 -> 74; 4251 -> 3711 | Mixed |
| use ReLU instead of tanh actor nonlinearities (ablation) | CEM-TD3 · HalfCheetah-v2, Hopper-v2, Walker2d-v2, Ant-v2 | mean return: 10725 -> 10267; 3613 -> 3604; 4711 -> 4311; 4251 -> 2264 | Hurt |
| use ReLU instead of tanh actor nonlinearities (ablation) | CEM-TD3 · Swimmer-v2 | mean return: 75 -> 118 | Helped |
| use ReLU instead of tanh actor nonlinearities (ablation) | CEM · Swimmer-v2 | performance: 60% performance drop | Hurt |
Closest papers by meaning
- The Ladder in Chaos: Improving Policy Learning by Harnessing the Parameter Evolving Path in A Low-dimensional Space (NeurIPS 2024 · similarity 0.89)
- EvoRainbow: Combining Improvements in Evolutionary Reinforcement Learning for Policy Search (ICML 2024 · similarity 0.88)
- ERL-TD: Evolutionary Reinforcement Learning Enhanced with Truncated Variance and Distillation Mutation (AAAI 2024 · similarity 0.88)
- Towards General-Purpose Model-Free Reinforcement Learning (ICLR 2025 · similarity 0.87)
- Double Buffers CEM-TD3: More Efficient Evolution and Richer Exploration (AAAI 2024 · similarity 0.87)
- DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Monte Carlo Tree Search (ICLR 2026 · 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.