AI Paper Atlas

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

Domain Arithmetic: One-Shot VLA Adaptation under Environmental Shifts

Taewook Kang, Taeheon Kim, Donghyun Shin, Jonghyun Choi

ECCV 2026 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Conditional distribution alignment · 1 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 domain vector to the base policypi0.5 · LIBERO novel viewpointssuccess rate: 24.6 percentage pointsImproved
DART instead of FLA (adapt whole model, not vision encoder)pi0.5 · LIBERO novel viewpointsaverage success rate: 74.3 -> 79.1Improved
DART instead of RETAINpi0.5 · LIBERO novel viewpointsaverage success rate: 69.6 -> 79.1Improved
one-shot fine-tuning instead of zero-shotpi0.5 · LIBERO novel viewpointsaverage success rate: 54.5 -> 31.5Hurt
DART instead of best baseline on real robotpi0.5 · Real-world UR10e novel viewpointaverage success rate: 55.0 -> 81.7Improved
DART instead of zero-shot on pi0-FASTpi0-FAST · LIBERO novel viewpointsaverage success rate: 73.4 -> 79.4Improved
DART instead of zero-shot cross-embodimentpi0.5 · MimicGen Panda to UR5e (cross-embodiment)average success rate: 62.0 -> 69.4Improved
add subspace alignment (filtering and scaling) (ablation)DART · LIBERO novel viewpointssuccess rateImproved
DART without subspace alignment instead of one-shot fine-tuning (ablation)DART w/o SA · LIBERO novel viewpointssuccess rateImproved
randomized SVD (r=256) instead of full SVD (ablation)DART · LIBERO novel viewpointsaverage success rate: 79.1 -> 78.7No clear effect
randomized SVD (r=256) instead of full SVD (ablation)DART · LIBERO novel viewpointsruntime: 15m 35s -> 6m 33sImproved
increase scaling coefficient alpha too far (ablation)DART · LIBERO novel viewpointssuccess rateHurt

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