Research map / Label-efficient and robust learning / Domain adaptation and generalization
The l2,1-Norm Stacked Robust Autoencoders for 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 |
|---|---|---|---|
| use l2,1-SRA instead of mSDA | l2,1-SRA · Amazon review AVG | accuracy (%): 82.84 -> 83.11 | Helped |
| use l2,1-SRA instead of mSDA | l2,1-SRA · Spam AVG | accuracy (%): 92.69 -> 94.15 | Helped |
| use l2,1-SRA instead of mSDA | l2,1-SRA · 20 newsgroups AVG | accuracy (%): 87.17 -> 90.86 | Helped |
| use l2,1-SRA instead of mSDA on Books to Electronics-style pairs where mSDA wins | l2,1-SRA · Amazon review D →E | accuracy (%): 85.06 -> 83.59 | Hurt |
| l2,1-SRA vs raw tf-idf SVM baseline | l2,1-SRA · Amazon review AVG | accuracy (%): 77.30 -> 83.11 | Helped |
| l2,1-SRA vs raw tf-idf SVM baseline | l2,1-SRA · Spam AVG | accuracy (%): 68.95 -> 94.15 | Helped |
| l2,1-SRA vs CODA co-training | l2,1-SRA · 20 newsgroups AVG | accuracy (%): 79.42 -> 90.86 | Helped |
| l2,1-SRA vs CODA on B →E | l2,1-SRA · Amazon review B →E | accuracy (%): 81.64 -> 76.79 | Hurt |
| remove nonlinear squashing step (ablation) | l2,1-SRA · Amazon/spam/20NG | accuracy (%) | Hurt |
| increase alpha to 1 (weak squashing) (ablation) | l2,1-SRA · B →D, Public →U0, Comp vs. Rec | accuracy (%) | Hurt |
| vary lambda (ablation) | l2,1-SRA · B →D, Public →U0, Comp vs. Rec | accuracy (%) | No clear effect |
| use 3-5 stacked layers instead of fewer or more (ablation) | l2,1-SRA · B →D, Public →U0, Comp vs. Rec | accuracy (%) | Helped |
Closest papers by meaning
- Adversarial Training Based Multi-Source Unsupervised Domain Adaptation for Sentiment Analysis (AAAI 2020 · similarity 0.86)
- Domain Aggregation Networks for Multi-Source Domain Adaptation (ICML 2020 · similarity 0.84)
- Multi-Adversarial Domain Adaptation (AAAI 2018 · similarity 0.84)
- Low-Rank Adapting Models for Sparse Autoencoders (ICML 2025 · similarity 0.83)
- Understanding and Improving Source-free Domain Adaptation from a Theoretical Perspective (CVPR 2024 · similarity 0.82)
- Proxy Denoising for Source-Free Domain Adaptation (ICLR 2025 · similarity 0.82)
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.