Research map / Label-efficient and robust learning / Domain adaptation and generalization
CyCADA: Cycle-Consistent Adversarial 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 feature-level adaptation to pixel-level adaptation (ablation) | CyCADA · SVHN → MNIST | accuracy: 70.3 ± 0.2 -> 90.4 ± 0.4 | Helped |
| Add feature-level adaptation to pixel-level adaptation (ablation) | CyCADA · USPS → MNIST | accuracy: 96.4 ± 0.1 -> 96.5 ± 0.1 | No clear effect |
| Add feature-level adaptation to pixel-level adaptation (ablation) | CyCADA · MNIST → USPS | accuracy: 95.6 ± 0.2 -> 95.6 ± 0.2 | No clear effect |
| Add pixel-level adaptation to feature-level adaptation (ablation) | CyCADA with VGG16-FCN8s · GTA5 → Cityscapes | mIoU: 29.2 -> 35.4 | Helped |
| Add feature-level adaptation to pixel-level adaptation (ablation) | CyCADA with DRN-26 · GTA5 → Cityscapes | mIoU: 37.0 -> 39.5 | Helped |
| Remove semantic consistency (ablation) | CyCADA · SVHN → MNIST | accuracy | Hurt |
| Remove cycle consistency (ablation) | CyCADA · SVHN → MNIST | semantic preservation | Hurt |
| Use CyCADA pixel-only instead of source-only training | FCN8s · SYNTHIA Fall → Winter | mIoU: 49.8 -> 63.3 | Helped |
| Use CyCADA pixel+feat instead of source-only training | VGG16-FCN8s · GTA5 → Cityscapes | mIoU: 17.9 -> 35.4 | Helped |
Closest papers by meaning
- Generate to Adapt: Aligning Domains Using Generative Adversarial Networks (CVPR 2018 · similarity 0.86)
- Visual Domain Adaptation by Consensus-Based Transfer to Intermediate Domain (AAAI 2020 · similarity 0.86)
- GCAN: Graph Convolutional Adversarial Network for Unsupervised Domain Adaptation (CVPR 2019 · similarity 0.86)
- Adversarial Discriminative Domain Adaptation (CVPR 2017 · similarity 0.86)
- Taking a Closer Look at Domain Shift: Category-Level Adversaries for Semantics Consistent Domain Adaptation (CVPR 2019 · similarity 0.86)
- CMDA: Cross-Modal and Domain Adversarial Adaptation for LiDAR-Based 3D Object Detection (AAAI 2024 · similarity 0.85)
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.