Research map / Reinforcement learning / Visual and representation RL
Correlation Priors for 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 |
|---|---|---|---|
| Model correlations between state-conditioned action distributions with a PG prior instead of independent Dirichlet priors | Imitation learning · Grid world | policy reconstruction (Hellinger distance and value loss) | Helped |
| Model correlations between state-conditioned transition distributions with a PG prior instead of independent Dirichlet priors | System identification · Grid world | transition estimation error (Hellinger distance) | Helped |
| Use correlated PG transition priors for posterior sampling | Bayesian reinforcement learning · Grid world | transitions needed to learn an effective strategy | Helped |
| Use correlated PG transition priors rather than correlation-agnostic priors | Greedy Bayesian reinforcement learning · Two-server queueing network | expected return | Helped |
| Model correlated subgoal distributions across states rather than using an uncorrelated subgoal model | Subgoal modeling · Grid world with walls | policy reconstruction (Hellinger distance) | Helped |
| Represent behavior through correlated subgoals rather than interpolating actions directly | Subgoal modeling · Grid world with walls | policy reconstruction (Hellinger distance) | Helped |
| Use correlated subgoal distributions to propagate goal information to unvisited states | Subgoal modeling · Grid world with walls | generalization to unvisited states | Helped |
Closest papers by meaning
- Exploiting Causal Graph Priors with Posterior Sampling for Reinforcement Learning (ICLR 2024 · similarity 0.86)
- Probabilistic Subgoal Representations for Hierarchical Reinforcement Learning (ICML 2024 · similarity 0.86)
- Bayesian Exploration Networks (ICML 2024 · similarity 0.85)
- Causal Imitation for Markov Decision Processes: a Partial Identification Approach (NeurIPS 2024 · similarity 0.85)
- Maximum Total Correlation Reinforcement Learning (ICML 2025 · similarity 0.85)
- Learning Task Belief Similarity with Latent Dynamics for Meta-Reinforcement Learning (ICLR 2025 · 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.