Research map / Label-efficient and robust learning / Domain adaptation and generalization
Importance Weighted Adversarial Nets for Partial Domain Adaptation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Add importance weighting of source samples using a first domain classifier | ADDA-grl-style adversarial adaptation · Office-Caltech partial adaptation (average) | accuracy (%): 92.31 -> 93.85 | Helped |
| Add target entropy minimization (ablation) | Proposed method · Office-Caltech partial adaptation (average) | accuracy (%): 93.72 -> 93.85 | Helped |
| Add target entropy minimization (ablation) | Proposed method · Office partial adaptation (average) | accuracy (%): 86.73 -> 87.57 | Helped |
| Use unshared instead of shared feature extractors (ablation) | Adversarial domain adaptation · A31→W10 | accuracy (%): 71.5 -> 76.3 | Helped |
| Use importance-weighted adversarial adaptation | AlexNet+bottleneck · Caltech256→Office10 (average) | accuracy (%): 49.86 -> 84.14 | Helped |
| Apply ordinary adversarial marginal alignment without partial-class weighting | AlexNet+bottleneck · Office-Caltech partial adaptation (average) | accuracy (%): 87.14 -> 79.83 | Hurt |
Closest papers by meaning
- Adversarial Reweighting for Partial Domain Adaptation (NeurIPS 2021 · similarity 0.91)
- Re-Weighted Adversarial Adaptation Network for Unsupervised Domain Adaptation (CVPR 2018 · similarity 0.90)
- Learning to Transfer Examples for Partial Domain Adaptation (CVPR 2019 · similarity 0.89)
- Domain-Symmetric Networks for Adversarial Domain Adaptation (CVPR 2019 · similarity 0.89)
- Adversarial Discriminative Domain Adaptation (CVPR 2017 · similarity 0.88)
- A DIRT-T Approach to Unsupervised Domain Adaptation (ICLR 2018 · similarity 0.88)
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.