Research map / Reinforcement learning / Policy gradient and actor-critic
$\mathrm{SO}(2)$-Equivariant Reinforcement Learning
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 |
|---|---|---|---|
| Use an SO(2)-equivariant network instead of the competing baseline | Equivariant DQN · Block Pulling, Object Picking, Drawer Opening | discounted reward | Helped |
| Use an SO(2)-equivariant actor and critic instead of baseline architectures | Equivariant SAC · Block Pulling, Object Picking, Drawer Opening | discounted reward | Helped |
| Use an SO(2)-equivariant actor and critic | SACfD · Four challenging manipulation tasks | discounted reward | Helped |
| Use an SO(2)-equivariant actor and critic instead of rotation-augmented RAD or DrQ | SACfD · Manipulation tasks | discounted reward | Helped |
| Use an SO(2)-equivariant actor and critic | SACfD trained on a fixed orientation · Random-orientation evaluation | discounted reward | Helped |
| Add an equivariant network to the actor (ablation) | SACfD with a conventional CNN critic · Six SACfD experiments | discounted reward | Helped |
| Add an equivariant network to both actor and critic (ablation) | SACfD · Six SACfD experiments | discounted reward | Helped |
| Use C8 equivariance instead of C4 or C2 equivariance (ablation) | Equivariant SACfD · Manipulation tasks | discounted reward | Helped |
| Use equivariant architecture when the environment has fixed orientation rather than rotational symmetry (ablation) | Equivariant SACfD · Fixed-orientation Block Pulling and Drawer Opening | discounted reward | No clear effect |
| Add rotational buffer augmentation to rotation-augmented RAD and DrQ (ablation) | RAD and DrQ · Manipulation tasks | discounted reward | Hurt |
Closest papers by meaning
- Boosting Sample Efficiency and Generalization in Multi-agent Reinforcement Learning via Equivariance (NeurIPS 2024 · similarity 0.90)
- Hierarchical Equivariant Policy via Frame Transfer (ICML 2025 · similarity 0.89)
- ${\rm E}(3)$-Equivariant Actor-Critic Methods for Cooperative Multi-Agent Reinforcement Learning (ICML 2024 · similarity 0.88)
- ET-SEED: EFFICIENT TRAJECTORY-LEVEL SE(3) EQUIVARIANT DIFFUSION POLICY (ICLR 2025 · similarity 0.88)
- A Practical Guide for Incorporating Symmetry in Diffusion Policy (NeurIPS 2025 · similarity 0.86)
- Learning (Approximately) Equivariant Networks via Constrained Optimization (NeurIPS 2025 · 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.