Research map / Reinforcement learning / Policy gradient and actor-critic
Exploration via Epistemic Value Estimation
Simon Schmitt, John Shawe-Taylor, Hado van Hasselt
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 posterior-sampled Thomson Sampling for action selection instead of ε-greedy (ablation) | Epistemic Q-Learning · Bsuite exploration tasks | exploration score | Improved |
| Bootstrap from a posterior-sampled Q-function instead of the Q-function at the MLE (ablation) | Epistemic Q-Learning · Bsuite exploration tasks | exploration score | Improved |
| Use the distributional update when estimating the Fisher instead of a regular non-distributional Q-Learning update (ablation) | Epistemic Q-Learning · Bsuite exploration tasks | exploration score | Improved |
| Use Leaky-ReLU activations instead of ReLU (ablation) | Epistemic Q-Learning · Bsuite exploration tasks | exploration score | Improved |
| Use Epistemic Q-Learning rather than DQN with ε-greedy exploration | Q-Learning · Deep Sea (30×30) | treasure discovery | Improved |
| Use Epistemic Q-Learning rather than Bootstrapped DQN | Bsuite agents · Bsuite | aggregate performance | No clear effect |
Closest papers by meaning
- Efficient Exploration via Epistemic-Risk-Seeking Policy Optimization (ICML 2023 · similarity 0.91)
- EUBRL: Epistemic Uncertainty Directed Bayesian Reinforcement Learning (ICLR 2026 · similarity 0.90)
- Epistemic Monte Carlo Tree Search (ICLR 2025 · similarity 0.88)
- Model-Value Inconsistency as a Signal for Epistemic Uncertainty (ICML 2022 · similarity 0.88)
- Beyond Optimism: Exploration With Partially Observable Rewards (NeurIPS 2024 · similarity 0.87)
- Bayesian Exploration Networks (ICML 2024 · 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.
