Research map / Reinforcement learning / Visual and representation RL
Bootstrap Latent-Predictive Representations for Multitask Reinforcement Learning
Zhaohan Daniel Guo, Bernardo Avila Pires, Bilal Piot, Jean-Bastien Grill, Florent Altché, Remi Munos, Mohammad Gheshlaghi Azar
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 PBL auxiliary representation learning instead of other representation-learning tasks | PopArt-IMPALA · DMLab-30 | mean capped human normalized score | Improved |
| Use PBL auxiliary representation learning instead of other representation-learning tasks | PopArt-IMPALA · Atari-57 | median human normalized score | Improved |
| Extend PBL forward-prediction horizon (ablation) | PBL · DMLab-30 | mean capped human normalized score | Improved |
| Extend prediction horizon with reverse prediction disabled, leaving random-projection latents (ablation) | PBL without reverse prediction · DMLab-30 | mean capped human normalized score | No clear effect |
| Disable reverse prediction and use random-projection latents (ablation) | PBL · DMLab-30 | mean capped human normalized score | Hurt |
| Add a random-projection latent prediction alongside the learned latent (ablation) | PBL · DMLab-30 | mean capped human normalized score | No clear effect |
| Use the larger network architecture (ablation) | PopArt-IMPALA · DMLab-30 | mean capped human normalized score | Improved |
| Use learned, meaningful future latent embeddings instead of random projections (ablation) | PBL · DMLab-like object-position probe | object position prediction loss | Improved |
Closest papers by meaning
- Bootstrapped Representations in Reinforcement Learning (ICML 2023 · similarity 0.89)
- TD-JEPA: Latent-predictive Representations for Zero-Shot Reinforcement Learning (ICLR 2026 · similarity 0.88)
- Representation Learning Enables Scalable Multitask Deep Reinforcement Learning (NeurIPS 2026 · similarity 0.87)
- Bridging State and History Representations: Understanding Self-Predictive RL (ICLR 2024 · similarity 0.86)
- Probabilistic Subgoal Representations for Hierarchical Reinforcement Learning (ICML 2024 · similarity 0.85)
- Towards Principled Representation Learning from Videos for Reinforcement Learning (ICLR 2024 · 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.
