Research map / Reinforcement learning / Policy gradient and actor-critic
Frictional Q-Learning
Hyunwoo Kim, Hyo Kyung Lee
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 total correlation term to the cVAE (ablation) | FQL · Ant-v4 | average return: 3204.6 -> 5102.1 | Improved |
| Add total correlation term to the cVAE (ablation) | FQL · Hopper-v4 | average return: 3370.5 -> 3370.5 | No clear effect |
| Use latent dimension 2 dim(A) instead of 1 (ablation) | FQL · Walker2d-v4 | average return: 4201.0 -> 4600.2 | Improved |
| Use latent dimension 2 dim(A) instead of 1 (ablation) | FQL · HalfCheetah-v4 | average return: 10096.0 -> 11429.0 | Improved |
| Use argmin Q background selection instead of uniform sampling (ablation) | FQL · Ant-v4 | average return: 3160.1 -> 5102.1 | Improved |
| Use argmin Q background selection instead of uniform sampling (ablation) | FQL · Walker2d-v4 | average return: 4729.1 -> 4600.2 | Hurt |
| Use replay buffer 1M instead of 0.1M (ablation) | FQL · MuJoCo (5 envs) | average return: Walker2d 2965.0 -> 4600.2; HalfCheetah 11939.1 -> 11429.0 | Mixed |
| Vary beta from 0 to 2 (ablation) | FQL · Hopper-v4 | average return: 2973.0 -> 3370.5 | Improved |
| Vary beta from 0 to 1 (ablation) | FQL · HalfCheetah-v4 | average return: 13497.0 -> 15969.6 | Improved |
| Use FQL instead of SAC | SAC · Humanoid-v4 | average return: 5682.62 -> 6437.15 | Improved |
| Use FQL instead of TD3 | TD3 · Hopper-v4 | average return: 3371.87 -> 3370.51 | No clear effect |
| Use FQL instead of TD3 | TD3 · Ant-v4 | average return: 6542.34 -> 6176.60 | No clear effect |
Closest papers by meaning
- Penalizing Infeasible Actions and Reward Scaling in Reinforcement Learning with Offline Data (ICML 2025 · similarity 0.86)
- Adaptive Neighborhood-Constrained Q Learning for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.85)
- From Weak Data to Strong Policy: Q-Targets Enable Provable In-Context Reinforcement Learning (NeurIPS 2026 · similarity 0.85)
- Actor-Free Continuous Control via Structurally Maximizable Q-Functions (NeurIPS 2025 · similarity 0.85)
- Learning from Suboptimal Data in Continuous Control via Auto-Regressive Soft Q-Network (ICML 2025 · similarity 0.85)
- Entropy-regularized Diffusion Policy with Q-Ensembles for Offline Reinforcement Learning (NeurIPS 2024 · 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.
