Research map / Reinforcement learning / Policy gradient and actor-critic
The Choice Function Framework for Online Policy Improvement
Murugeswari Issakkimuthu, Alan Fern, Prasad Tadepalli
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 |
|---|---|---|---|
| apply best LDCF online search on top of linear base policy | Policy 1 (linear) · Game-of-Life problem 1 | normalized reward: 2.57 | Improved |
| apply best LDCF online search on top of linear base policy | Policy 1 (linear) · Game-of-Life problem 10 | normalized reward: 2.11 | Improved |
| apply best LDCF online search on top of linear base policy | Policy 1 (linear) · Game-of-Life problem 6 | normalized reward: 1.05 | Improved |
| apply best LDCF online search on top of non-linear base policy | Policy 2 (3 hidden layers) · Game-of-Life problem 10 | normalized reward: 1.23 | Improved |
| apply best LDCF online search on top of non-linear base policy | Policy 2 (3 hidden layers) · Game-of-Life problems 2, 3, 6, 9 | normalized reward: 0.95 | Hurt |
| apply best LDCF online search on top of non-linear base policy | Policy 2 (3 hidden layers) · Game-of-Life problem 5 | normalized reward: 1.00 | No clear effect |
| use stronger base policy (non-linear instead of linear) | LDCF online search · Game-of-Life problem 1 | normalized reward gain: 2.57 -> 1.08 | Hurt |
| allow longer decision time per step | LDCF configurations · Game-of-Life | normalized reward | Mixed |
| small FSSS sampling width C=3 on highly stochastic problems | LDCF with FSSS · Game-of-Life stochastic problems | normalized reward | Hurt |
Closest papers by meaning
- Learning Optimal Advantage from Preferences and Mistaking It for Reward (AAAI 2024 · similarity 0.86)
- Online Markov Decision Processes Configuration with Continuous Decision Space (AAAI 2024 · similarity 0.85)
- On the Convergence of No-Regret Dynamics in Information Retrieval Games with Proportional Ranking Functions (ICLR 2025 · similarity 0.85)
- Design Considerations in Offline Preference-based RL (ICML 2025 · similarity 0.85)
- Policy-labeled Preference Learning: Is Preference Enough for RLHF? (ICML 2025 · similarity 0.85)
- Towards Theoretical Understanding of Sequential Decision Making with Preference Feedback (ICML 2025 · similarity 0.85)
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.
