AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Transport or Discard: Robust Unbalanced Optimal Transport for Cross-Domain Policy Adaptation

Wenyu Chen, Yujia Zhang, Wei Guo, Linli Ma, Yanbo Wang, Pinle Qin, Jianchao Zeng

ICML 2026 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Mixture and distillation for domain shift

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
replace balanced OT filtering (OTDF) with robust unbalanced OT (ROOT)OTDF · D4RL kinematic shift (36 tasks)total normalized score: 1547.6 -> 2073.5Improved
use ROOT instead of strongest baseline DmCDmC · D4RL kinematic shift (36 tasks)total normalized score: 1902.2 -> 2073.5Improved
use ROOT instead of OTDF on Hopper-medium to expert targetOTDF · Hopper-M to Expert (kinematic)normalized score: 35.0 -> 99.4Improved
use ROOT instead of MOBODYMOBODY · ODRL gravity/friction shifts (32 settings)total normalized score: 1427.26 -> 1477.97Improved
add threshold-based hard filtering on top of ROOT weights (tau_th 0.0 to 2.0) (ablation)ROOT · halfcheetah-gravity, halfcheetah-morph, hopper-kinematic, walker2d-morphaverage returnNo clear effect
vary source-side relaxation lambda_src (0.01 to 1.0) (ablation)ROOT · same four environmentsaverage returnMixed
vary target-side relaxation lambda_tar in [0.05, 2.0] (ablation)ROOT · same four environmentsaverage returnNo clear effect

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