Research map / Label-efficient and robust learning / Domain adaptation and generalization
Bi-Directional Generation for Unsupervised 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 MMD to single-direction target-to-source generation (ablation) | BDG Variant 1 · Office-31 mean accuracy | accuracy (%): 81.1 -> 87.4 | Helped |
| Add MMD to single-direction target-to-source generation (ablation) | BDG Variant 1 · Office-Home mean accuracy | accuracy (%): 61.0 -> 67.8 | Helped |
| Add MMD and consistency loss to GAN losses (ablation) | BDG · Office-31 mean accuracy | accuracy (%): 88.3 -> 88.5 | Helped |
| Add MMD and consistency loss to GAN losses (ablation) | BDG · Office-Home mean accuracy | accuracy (%): 68.2 -> 68.7 | Helped |
| Add MMD to GAN and consistency losses (ablation) | BDG Variant 4 · Office-Home mean accuracy | accuracy (%): 63.5 -> 68.7 | Helped |
| Use bi-directional rather than single-direction generation (ablation) | BDG · Office-31 mean accuracy | accuracy (%): 81.1 -> 88.5 | Helped |
| Use bi-directional rather than single-direction generation (ablation) | BDG · Office-Home mean accuracy | accuracy (%): 61.0 -> 68.7 | Helped |
| Use BDG instead of TADA | Office-31 · Office-31 mean accuracy | accuracy (%): 88.4 -> 88.5 | Helped |
| Use BDG instead of SymNets or TADA | Office-Home · Office-Home mean accuracy | accuracy (%): 67.6 -> 68.7 | Helped |
| Use BDG instead of the strongest baseline | Office-31 A→W · Office-31 (A→W) | accuracy (%): 94.3±0.3 -> 93.6±0.4 | Hurt |
| Use BDG instead of the strongest baseline | Office-31 D→A · Office-31 (D→A) | accuracy (%): 74.6±0.6 -> 73.2±0.2 | Hurt |
Closest papers by meaning
- A Prototype-Oriented Framework for Unsupervised Domain Adaptation (NeurIPS 2021 · similarity 0.90)
- Visual Domain Adaptation by Consensus-Based Transfer to Intermediate Domain (AAAI 2020 · similarity 0.90)
- Co-regularized Alignment for Unsupervised Domain Adaptation (NeurIPS 2018 · similarity 0.89)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (ICML 2025 · similarity 0.88)
- Invertible Projection and Conditional Alignment for Multi-Source Blended-Target Domain Adaptation (AAAI 2025 · similarity 0.88)
- Gradient-Based Sample Selection for Black-Box Universal Domain Adaptation (AAAI 2025 · 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.