Research map / Label-efficient and robust learning / Domain adaptation and generalization
Deep Reconstruction-Classification Networks for Unsupervised Domain Adaptation
Muhammad Ghifary, W. Bastiaan Kleijn, Mengjie Zhang, David Balduzzi, Wen Li
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 changed | Where | Measured | Effect |
|---|---|---|---|
| reconstruct unlabeled target images only rather than unlabeled source images (DRCN vs DRCN_s) (ablation) | DRCN · SVHN->MNIST | accuracy (%): 73.66 -> 81.97 | Improved |
| reconstruct unlabeled target images only rather than source+target (DRCN vs DRCN_st) (ablation) | DRCN · SVHN->MNIST | accuracy (%): 75.88 -> 81.97 | Improved |
| reconstruct target only vs source only (ablation) | DRCN · USPS->MNIST | accuracy (%): 65.96 -> 73.67 | Improved |
| reconstruct target only vs source+target (ablation) | DRCN · USPS->MNIST | accuracy (%): 68.64 -> 73.67 | Improved |
| reconstruct target only vs source only (ablation) | DRCN · MNIST->SVHN | accuracy (%): 34.29 -> 40.05 | Improved |
| reconstruct target only vs source+target (ablation) | DRCN · STL->CIFAR | accuracy (%): 55.26 -> 58.86 | Improved |
| reconstruct target only vs source+target (ablation) | DRCN · MNIST->USPS | accuracy (%): 91.15 -> 91.80 | No clear effect |
| reconstruct target only vs source only (ablation) | DRCN · CIFAR->STL | accuracy (%): 63.02 -> 66.37 | Improved |
| replace adversarial domain-confusion training (ReverseGrad) with joint reconstruction training (DRCN) | DRCN · SVHN->MNIST | accuracy (%): 73.91 -> 81.97 | Improved |
| replace ReverseGrad with DRCN | DRCN · MNIST->SVHN | accuracy (%): 35.67 -> 40.05 | Improved |
| replace ReverseGrad with DRCN | DRCN · USPS->MNIST | accuracy (%): 74.01 -> 73.67 | Hurt |
| joint reconstruction-classification training instead of greedy pretraining then finetuning (SCAE_t) | DRCN · SVHN->MNIST | accuracy (%): 65.57 -> 81.97 | Improved |
Closest papers by meaning
- Deep Residual Correction Network for Partial Domain Adaptation (TPAMI 2020 · similarity 0.90)
- Learning Transferrable Representations for Unsupervised Domain Adaptation (NeurIPS 2016 · similarity 0.89)
- Unsupervised Domain Adaptation with Residual Transfer Networks (NeurIPS 2016 · similarity 0.89)
- Image to Image Translation for Domain Adaptation (CVPR 2018 · similarity 0.88)
- Improving Domain-Specific Classification by Collaborative Learning with Adaptation Networks (AAAI 2019 · similarity 0.87)
- Beyond Sharing Weights for Deep Domain Adaptation (TPAMI 2018 · similarity 0.87)
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.
