AI Paper Atlas

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

Keypoint-Guided Optimal Transport with Applications in Heterogeneous Domain Adaptation

Xiang Gu, Yucheng Yang, Wei Zeng, Jian Sun, Zongben Xu

NeurIPS 2022 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Statistical alignment and hypothesis testing · 6 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
use relation-preserving guidance instead of distance-preserving guidance (ablation)KPG-RL (SH) · Office-31 HDAaccuracy: 64.9 -> 79.9Improved
solve with Sinkhorn instead of linear programming (ablation)KPG-RL · Office-31 HDAaccuracy: 79.3 -> 79.9No clear effect
use GW-based variant instead of Sinkhorn variant (ablation)KPG-RL · Office-31 HDAaccuracy: 79.9 -> 79.6No clear effect
add mask-based constraint on transport plan (ablation)GW · Office-31 HDAaccuracy: 35.6 -> 60.8Improved
add mask-based constraint on transport plan (ablation)HOT · Office-31 HDAaccuracy: 44.3 -> 55.9Improved
add mask-based constraint on transport plan (ablation)TLB · Office-31 HDAaccuracy: 35.7 -> 58.3Improved
plain GW/HOT/TLB transport vs labeled target onlyGW · Office-31 HDAaccuracy: 60.6 -> 35.6Hurt
relation-preserving KPG-RL vs class-center-preserving SGWKPG-RL (SH) · Office-31 HDAaccuracy: 67.3 -> 79.9Improved
add KPG-RL transport with 1 labeled target shot per classLabeled-target-only · Office-31 HDA 1-shotaccuracy: 60.6 -> 79.9Improved
add KPG-RL transport with 3 labeled target shots per classLabeled-target-only · Office-31 HDA 3-shotaccuracy: 80.9 -> 87.7Improved
more labeled target shots (1 to 3) shrinks gain of transferKPG-RL (SH) · Office-31 HDAaccuracy: 79.9 -> 87.7Mixed
add partial-KPG-RL detection of unknown classes over rejection baselineBaseline · Office-31 open-set HDAHOS: 48.7 -> 69.1Improved

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