Research map / Reinforcement learning / Visual and representation RL
Language as an Abstraction for Hierarchical Deep 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 compositional language instead of lossless non-compositional latent instruction representation (ablation) | low-level instruction-following policy · CLEVR-Robot | instructions achieved per episode | Helped |
| use language instead of one-hot instruction encoding with many instructions (ablation) | low-level policy · CLEVR-Robot | instructions achieved per episode | Mixed |
| add hindsight instruction relabeling (HIR) (ablation) | low-level policy · CLEVR-Robot | instructions achieved per episode | Helped |
| use sequential language encoder instead of bag-of-words (ablation) | low-level policy · CLEVR-Robot | final instructions per episode | Helped |
| hierarchical language abstraction (HAL) vs flat DDQN, HIRO, Option-Critic | high-level tasks · CLEVR-Robot tasks (a-c) | success | Helped |
| systematic (zero-shot) split instead of standard split | Language low-level policy · CLEVR-Robot | test instructions per episode: 21.49 -> 8.13 | Hurt |
Closest papers by meaning
- Prompted Policy Search: Reinforcement Learning through Linguistic and Numerical Reasoning in LLMs (NeurIPS 2025 · similarity 0.87)
- Text2Reward: Reward Shaping with Language Models for Reinforcement Learning (ICLR 2024 · similarity 0.86)
- Ground-Compose-Reinforce: Grounding Language in Agentic Behaviours using Limited Data (NeurIPS 2025 · similarity 0.86)
- RLZero: Direct Policy Inference from Language Without In-Domain Supervision (NeurIPS 2025 · similarity 0.85)
- Divide and Conquer: Grounding LLMs as Efficient Decision-Making Agents via Offline Hierarchical Reinforcement Learning (ICML 2025 · similarity 0.85)
- Sketch Then Paint: Hierarchical Reinforcement Learning for Diffusion Multi-Modal Large Language Models (NeurIPS 2026 · similarity 0.85)
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.