AI Paper Atlas

Research map / Reinforcement learning / Offline RL

Lightweight Online Reinforcement Learning for Block Decomposition of CAD Models

Xitong Luo, Zhenghan Wu, Yucong Wang, Yi Cai

ECCV 2026 · Reinforcement learning › Offline RL · approach: Diffusion models for offline decision making · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
add frozen pre-trained feature encoder instead of joint training (ablation)Online SAC framework · Self-built datasetTotal time (min): 39.17 -> 34.12Improved
add frozen feature encoder instead of joint training (ablation)Online SAC framework · Public dataset subsetSVR (%): 86.23 -> 90.14Improved
add frozen feature encoder instead of joint training (ablation)Online SAC framework · Public dataset subsetRL time (min): 4.51 -> 2.38Improved
add frozen encoder (ablation)Online SAC framework · Self-built datasetSVR (%): 100.00 -> 100.00No clear effect
add continuous-to-discrete action projection (ablation)Online SAC framework · Self-built datasetSVR (%): 86.39 -> 100.00Improved
add continuous-to-discrete action projection (ablation)Online SAC framework · Public dataset subsetQHMR (%): 72.23 -> 94.96Improved
add action projection (ablation)Online SAC framework · Self-built datasetMin_steps: 26.37 -> 14.04Improved
add action projection (ablation)Online SAC framework · Self-built datasetRL time (min): 2.44 -> 2.25Hurt
pure online RL instead of offline RLOnline SAC framework · Cross-paradigmSVR (%): 90.01 -> 95.79Improved
pure online RL instead of offline RLOnline SAC framework · Cross-paradigmTotal delivery time (min): 138.09 -> 36.05Improved
drop offline pre-training relative to Hybrid-FTOnline SAC framework · Cross-paradigmSVR (%): 96.10 -> 95.79No clear effect
drop offline pre-training relative to Hybrid-FTOnline SAC framework · Cross-paradigmTotal delivery time (min): 173.65 -> 36.05Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy