Research map / Reinforcement learning / Policy gradient and actor-critic
Successor Uncertainties: Exploration and Uncertainty in Temporal Difference Learning
David Janz, Jiri Hron, Przemysław Mazur, Katja Hofmann, José Miguel Hernández-Lobato, Sebastian Tschiatschek
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 Successor Uncertainties instead of Bootstrapped DQN | Atari 2600 agent · Atari 2600 (49 games) | median human normalised score: 1.60 -> 2.09 | Improved |
| use Successor Uncertainties instead of UBE | Atari 2600 agent · Atari 2600 (49 games) | median human normalised score: 1.07 -> 2.09 | Improved |
| use Successor Uncertainties instead of DQN with epsilon-greedy | Atari 2600 agent · Atari 2600 (49 games) | superhuman performance %: 48.98% -> 77.55% | Improved |
| use factorised-posterior UBE or Bootstrapped DQN instead of SU on randomised binary tree | exploration algorithm · binary tree MDP | episodes to learn optimal policy | Hurt |
| use SU instead of Bootstrapped DQN on binary tree | exploration algorithm · binary tree MDP | computation to reach similar performance: approximately 25 times less computation | Improved |
| SU vs Bootstrapped DQN on individual games Breakout and Montezuma-type hard games | SU · Atari raw scores | score: 352.7 vs 855.0 (Breakout) | Mixed |
Later papers that compared with this method
1 later accepted paper reports Successor Uncertainties as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Posterior Sampling for Deep Reinforcement Learning (ICML 2023 · better on 4 of 4 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Propagating Uncertainty in Reinforcement Learning via Wasserstein Barycenters (NeurIPS 2019 · similarity 0.89)
- Posterior Sampling for Deep Reinforcement Learning (ICML 2023 · similarity 0.89)
- Randomized Exploration in Reinforcement Learning with General Value Function Approximation (ICML 2021 · similarity 0.88)
- Universal Value-Function Uncertainties (ICLR 2026 · similarity 0.88)
- Random Latent Exploration for Deep Reinforcement Learning (ICML 2024 · similarity 0.88)
- Parameterized Indexed Value Function for Efficient Exploration in Reinforcement Learning (AAAI 2020 · 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.
