Research map / Reinforcement learning / Visual and representation RL
Incremental Reinforcement Learning with Dual-Adaptive ε-Greedy Exploration
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 |
|---|---|---|---|
| Use DAE instead of Rainbow exploration | Rainbow · Incremental Atari | mean best normalized score: 3.23 -> 6.11 | Helped |
| Use DAE instead of Rainbow exploration | Rainbow · Incremental Atari | median best normalized score: 2.11 -> 3.97 | Helped |
| Train with expanding state and action spaces rather than a fixed environment | Rainbow · Atari benchmark | mean best normalized score: 35.15% | Hurt |
| Use DAE rather than Rainbow on Incremental Atari | Rainbow · Incremental Atari | normalized score: 88.76% | Helped |
| Use DAE rather than retraining from scratch | DQN on Expanding World · Expanding World | training steps to reach near-optimal policies: 76.37% less | Helped |
| Use DAE rather than the best baseline, ε-greedy | DQN on Expanding World · Expanding World | training steps to reach near-optimal policies: 73.58% less | Helped |
| Explore with DAE's least-tried action selection | Exploration on MountainCar · MountainCar | state coverage | Helped |
| Explore with DAE's least-tried action selection | Exploration on GridWorld · GridWorld | state coverage | Helped |
| Explore with DAE's least-tried action selection | Exploration on Labyrinth · Labyrinth | exploration efficiency | Helped |
| Explore with DAE's least-tried action selection | Exploration on MultiRoom · MultiRoom | rooms reached | Helped |
Closest papers by meaning
- LLM-Explorer: A Plug-in Reinforcement Learning Policy Exploration Enhancement Driven by Large Language Models (NeurIPS 2025 · similarity 0.88)
- Beyond Optimism: Exploration With Partially Observable Rewards (NeurIPS 2024 · similarity 0.87)
- Exploration from a Primal-Dual Lens: Value-Incentivized Actor-Critic Methods for Sample-Efficient Online RL (NeurIPS 2025 · similarity 0.87)
- Decoupling Exploration and Policy Optimization: Uncertainty Guided Tree Search for Hard Exploration (NeurIPS 2026 · similarity 0.87)
- Exploring Starts Are Not Enough: Counterexamples and a Fix for Monte Carlo Exploring Starts (NeurIPS 2026 · similarity 0.87)
- DARLING: Detection Augmented Reinforcement Learning with Non-Stationary Guarantees (NeurIPS 2026 · similarity 0.87)
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.