AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

FocalPolicy: Frequency-Optimized Chunking and Locally Anchored Flow Matching for Coherent Visuomotor Policy

Qian He, Zhenshuo Yang, Wenqi Liang, Chunhui Hao, Nicu Sebe, Jiandong Tian

ICML 2026 · Reinforcement learning › Visual and representation RL · approach: Dexterous manipulation transfer · 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 Foresight Composite Objective and Locally Anchored Sampling to consistency flow policyFocalPolicy · Adroit and MetaWorld 53 taskssuccess rate (%): 79.9 -> 83.6 (vs DP3)Improved
add Locally Anchored Sampling (ablation)FocalPolicy · 24 tasks ablationaverage success rate (%): 64.3 -> 77.9Improved
add FCO on top of LAS (ablation)FocalPolicy · 24 tasks ablationaverage success rate (%): 54.1 -> 64.3Improved
use frequency-domain loss over multi-chunk instead of time-domain loss over long horizon (ablation)FCO · 24 tasks ablationaverage success rate (%): 61.1 -> 77.9Improved
add proximal time-domain loss alongside frequency-domain loss (ablation)FCO · 24 tasks ablationaverage success rate (%): 72.8 -> 77.9Improved
sample anchor time r stochastically near terminal instead of fixing r = 1 (ablation)LAS · 24 tasks ablationaverage success rate (%): 73.3 -> 77.9Improved
bias anchor time r toward terminal (mu 4.0, sigma 1.6) instead of near 0.6 (ablation)LAS · 7 tasksaverage success rate (%): 13.9 -> 69.7Improved
integrate FCO into DP3DP3 · 7 simulation tasksaverage success rate (%): 57.8 -> 60.6Improved
integrate LAS into FlowPolicyFlowPolicy · 7 simulation tasksaverage success rate (%): 55.4 -> 63.3Improved
aggregate N=3 chunks instead of N=5 (H=4) (ablation)FocalPolicy · 24 tasksaverage success rate (%): 64.6 -> 77.9Improved
increase action chunk size H from 4 to 8 or 12 (ablation)FocalPolicy · 24 tasksaverage success rate (%): 77.9 -> 70.3 (H=8, N=3)Hurt
use FocalPolicy instead of FlowPolicy for inter-chunk coherenceFocalPolicy · MetaWorld Sweep IntoAction Total Variation: 1.66 -> 1.45Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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