Research map / Reinforcement learning / Visual and representation RL
The Natural Language of Actions
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 |
|---|---|---|---|
| Represent the drawing state with summed Act2Vec action embeddings instead of one-hot or random embeddings | PPO square drawing · Quick, Draw! square drawing | drawing performance | Helped |
| L2-normalize Act2Vec embeddings (ablation) | PPO square drawing · Quick, Draw! square drawing | drawing performance | No clear effect |
| Use Act2Vec rather than one-hot or random state embeddings | PPO square drawing · Quick, Draw! square drawing | variance across trials | Helped |
| Use length-2 action sequences instead of length-1 actions with ε-uniform exploration (ablation) | Q-Embedding DQN · Seek-Avoid | total reward: 20 percent increase in total reward | Helped |
| Use length-3 action sequences with ε-uniform exploration (ablation) | Q-Embedding DQN · Seek-Avoid | total reward | No clear effect |
| Use cluster-based k-Exp for length-3 action sequences instead of uniform exploration (ablation) | Q-Embedding DQN · Seek-Avoid | total reward | Helped |
Closest papers by meaning
- Learning Action Representations for Reinforcement Learning (ICML 2019 · similarity 0.92)
- Simple Emergent Action Representations from Multi-Task Policy Training (ICLR 2023 · similarity 0.89)
- HyAR: Addressing Discrete-Continuous Action Reinforcement Learning via Hybrid Action Representation (ICLR 2022 · similarity 0.88)
- For SALE: State-Action Representation Learning for Deep Reinforcement Learning (NeurIPS 2023 · similarity 0.88)
- In-Context Reinforcement Learning for Variable Action Spaces (ICML 2024 · similarity 0.86)
- NS-VLA: Towards Neuro-Symbolic Vision-Language-Action Models (NeurIPS 2026 · 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.