Research map / Reinforcement learning / Visual and representation RL
Discrete Compositional Representations as an Abstraction for Goal Conditioned 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 |
|---|---|---|---|
| Discretize visual goal representations into factorial codes | Goal-DQN · LoopWorld and SpiralWorld, including novel goals | goal-reaching performance | Helped |
| Use more discrete factors, increasing G to 16 (ablation) | DGRL + Goal-DQN · LoopWorld | convergence speed | Helped |
| Add discrete factorial goal representations | AMIGO · MiniGrid KCharder (KeyCorrS4R3) | task-solving efficiency | Helped |
| Use G=4 factors instead of G=8 or G=16 (ablation) | DGRL + HRAC · KeyChest | sample efficiency | Hurt |
| Use G=8 or G=16 factors (ablation) | DGRL + HRAC · KeyChest | sample efficiency | Helped |
| Add discrete goal representations | HRAC · AntMazeSparse, AntFall, and AntPush | success rate | Helped |
| Add discrete goal representations | RIS · AntU and AntMaze navigation | sample efficiency | Helped |
| Use G=4 factors instead of G=8 or G=16 (ablation) | DGRL + RIS · Sawyer robotic manipulation | performance | Hurt |
| Add discrete bottleneck representations | RIS · Sawyer robotic manipulation | sample efficiency | Helped |
| Add VIB to DGRL representation learning (ablation) | Robot image representation · Robot images with background distractors | noise removal | Helped |
| Use an autoencoder with two discrete factors | Color-MNIST representation · Color-MNIST | factor separation | Helped |
Closest papers by meaning
- Compositional Automata Embeddings for Goal-Conditioned Reinforcement Learning (NeurIPS 2024 · similarity 0.90)
- GoalLadder: Incremental Goal Discovery with Vision-Language Models (NeurIPS 2025 · similarity 0.89)
- Flattening Hierarchies with Policy Bootstrapping (NeurIPS 2025 · similarity 0.88)
- Reconciling Spatial and Temporal Abstractions for Goal Representation (ICLR 2024 · similarity 0.88)
- Improving the Continuity of Goal-Achievement Ability via Policy Self-Regularization for Goal-Conditioned Reinforcement Learning (ICML 2025 · similarity 0.88)
- Diffusion Subgoal Planning for Long-Horizon Offline Goal-Conditioned Reinforcement Learning (NeurIPS 2026 · similarity 0.88)
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.