Research map / Reinforcement learning / Policy gradient and actor-critic
Keep Various Trajectories: Promoting Exploration of Ensemble Policies in Continuous Control
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 |
|---|---|---|---|
| ensemble TD3 sub-policies with trajectory-aware exploration (TEEN) instead of plain TD3 | TD3 · HalfCheetah-v3 1M | average return: 9759 -> 11914 | Helped |
| TEEN instead of TD3 | TD3 · Ant-v3 1M | average return: 5142 -> 5930 | Helped |
| TEEN instead of SAC | SAC · Ant-v3 500K | average return: 3302 -> 5009 | Helped |
| TEEN instead of SUNRISE ensemble | SUNRISE · HalfCheetah-v3 1M | average return: 6269 -> 11914 | Helped |
| UCB-based ensemble exploration (SUNRISE) instead of TD3 | TD3 · HalfCheetah-v3 1M | average return: 9759 -> 6269 | Hurt |
| TEEN instead of TD3 | TD3 · Fish-Swim 1M | average return: 367 -> 519 | Helped |
| TEEN instead of SAC | SAC · Walker-Run 500K | average return: 688 -> 754 | Helped |
| TEEN instead of SAC | SAC · Walker-Walk 1M | average return: 976 -> 974 | No clear effect |
| increase number of target values M (ablation) | TEEN · Ant-v3 | estimation bias | Hurt |
| increase ensemble size N (ablation) | TEEN · Ant-v3 | performance and underestimation | Helped |
| add trajectory-aware exploration term (alpha=0 -> alpha=0.2) (ablation) | TEEN · Ant-v3 | performance | Helped |
| increase weight alpha to large values (ablation) | TEEN · Ant-v3 | performance | Hurt |
Closest papers by meaning
- Dynamic Optimizations of LLM Ensembles with Two-Stage Reinforcement Learning Agents (ICML 2026 · similarity 0.88)
- Diverse Projection Ensembles for Distributional Reinforcement Learning (ICLR 2024 · similarity 0.88)
- The Curse of Diversity in Ensemble-Based Exploration (ICLR 2024 · similarity 0.87)
- Oracle-Efficient Reinforcement Learning for Max Value Ensembles (NeurIPS 2024 · similarity 0.87)
- PWM: Policy Learning with Multi-Task World Models (ICLR 2025 · similarity 0.86)
- Enhancing Off-Policy Constrained Reinforcement Learning through Adaptive Ensemble C Estimation (AAAI 2024 · similarity 0.86)
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.