Research map / Reinforcement learning / Visual and representation RL
REFUEL: Exploring Sparse Features in Deep Reinforcement Learning for Fast Disease Diagnosis
Yu-Shao Peng, Kai-Fu Tang, Hsuan-Tien Lin, Edward Chang
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 |
|---|---|---|---|
| add reward shaping and feature rebuilding (ablation) | REINFORCE · 200 diseases test | top-5 accuracy: 66.55 -> 79.68 | Improved |
| add reward shaping and feature rebuilding (ablation) | REINFORCE · 300 diseases test | top-5 accuracy: 31.18 -> 71.17 | Improved |
| add reward shaping and feature rebuilding (ablation) | REINFORCE · 400 diseases test | top-5 accuracy: 2.09 -> 67.09 | Improved |
| add reward shaping alone (ablation) | REINFORCE · 200/300/400 diseases training | training accuracy and reward | Improved |
| add feature rebuilding on top of reward shaping (ablation) | RESHAPE · 200/300/400 diseases training | training accuracy | Improved |
| add reward shaping (and rebuilding) to reduce repeated queries (ablation) | REINFORCE · 300/400 diseases training | repeat rate | Improved |
| use REFUEL instead of hierarchical model | hierarchical model of AAAI 2018 · 73 diseases | top-5 accuracy: 77.94 -> 91.71 | Improved |
| use REFUEL instead of hierarchical model | hierarchical model of AAAI 2018 · 255 diseases | top-1 accuracy: 26.26 -> 50.66 | Improved |
| REFUEL vs baseline average number of steps (ablation) | REINFORCE · 200 diseases | #Steps: 6.66 -> 8.01 | Hurt |
| REFUEL vs hierarchical model average steps (ablation) | hierarchical model · 73 diseases | #Steps: 7.15 -> 7.52 | Hurt |
Closest papers by meaning
- Context-Aware Symptom Checking for Disease Diagnosis Using Hierarchical Reinforcement Learning (AAAI 2018 · similarity 0.89)
- MTDiag: An Effective Multi-Task Framework for Automatic Diagnosis (AAAI 2023 · similarity 0.87)
- Highly Efficient Self-Adaptive Reward Shaping for Reinforcement Learning (ICLR 2025 · similarity 0.81)
- ORSO: Accelerating Reward Design via Online Reward Selection and Policy Optimization (ICLR 2025 · similarity 0.81)
- DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning (NeurIPS 2025 · similarity 0.81)
- Reward Shaping for Reinforcement Learning with An Assistant Reward Agent (ICML 2024 · similarity 0.81)
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.
