Research map / Reinforcement learning / Visual and representation RL
Principled Analysis of Deep Reinforcement Learning Evaluation and Design Paradigms
Ezgi Korkmaz
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 plain Dueling instead of higher-capacity distributional C51 | C51 · ALE 100K | human normalized median: 0.0941 -> 0.2304 | Improved |
| use plain Dueling instead of QR-DQN | QR-DQN · ALE 100K | human normalized median: 0.0820 -> 0.2304 | Improved |
| use plain Dueling instead of IQN | IQN · ALE 100K | human normalized median: 0.0528 -> 0.2304 | Improved |
| use plain Dueling instead of Double-Q | Double-Q · ALE 100K | human normalized mean: 0.3169 -> 0.2923 | Hurt |
| use plain Dueling instead of Prior | Prior · ALE 100K | 20th percentile: 0.0267 -> 0.0764 | Improved |
| compare DrQ (ICLR) against core Dueling baseline rather than DER | DrQ (ICLR) · ALE 100K | performance gain over baseline: 82% -> 11% | No clear effect |
| compare DrQ (NeurIPS reimplementation) to Dueling | Dueling · ALE 100K | performance: 15% better | Hurt |
Closest papers by meaning
- Stable Gradients for Stable Learning at Scale in Deep Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- Scaling Offline Evaluation of Reinforcement Learning Agents through Abstraction (AAAI 2024 · similarity 0.87)
- Position: Benchmarking is Limited in Reinforcement Learning Research (ICML 2024 · similarity 0.87)
- Analyzing Generalization in Policy Networks: A Case Study with the Double-Integrator System (AAAI 2024 · similarity 0.87)
- Efficient Reinforcement Learning by Discovering Neural Pathways (NeurIPS 2024 · similarity 0.86)
- Deep Reinforcement Learning from Hierarchical Preference Design (ICML 2025 · 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.
