Research map / Label-efficient and robust learning / Domain adaptation and generalization
Beyond Separability: Analyzing the Linear Transferability of Contrastive Representations to Related Subpopulations
Jeff Z. HaoChen, Colin Wei, Ananya Kumar, Tengyu Ma
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 |
|---|---|---|---|
| use class-mean classifier (PFA t=1) instead of linear probe | PFA on spectral contrastive features · BREEDS Living-17 target | accuracy (%): 54.7 -> 67.4 | Improved |
| use class-mean classifier (PFA t=1) instead of linear probe | PFA on spectral contrastive features · BREEDS Entity-30 target | accuracy (%): 46.4 -> 62.3 | Improved |
| increase t from 1 to 2 (add preconditioner) (ablation) | PFA on spectral contrastive features · BREEDS Living-17 target | accuracy (%): 67.4 -> 72.0 | Improved |
| increase t from 1 to 2 (add preconditioner) (ablation) | PFA on spectral contrastive features · BREEDS Entity-30 target | accuracy (%): 62.3 -> 65.1 | Improved |
| use class-mean classifier (PFA t=1) instead of linear probe | PFA on spectral contrastive features · BREEDS Living-17 source | accuracy (%): 91.3 -> 92.6 | Improved |
| use class-mean classifier (PFA t=1) instead of linear probe | PFA on spectral contrastive features · BREEDS Entity-30 source | accuracy (%): 84.8 -> 82.8 | Hurt |
| increase t from 1 to 2 (add preconditioner) (ablation) | PFA on spectral contrastive features · BREEDS Living-17 source | accuracy (%): 92.6 -> 90.5 | Hurt |
| increase t from 1 to 2 (add preconditioner) (ablation) | PFA on spectral contrastive features · BREEDS Entity-30 source | accuracy (%): 82.8 -> 77.3 | Hurt |
Closest papers by meaning
- Connect, Not Collapse: Explaining Contrastive Learning for Unsupervised Domain Adaptation (ICML 2022 · similarity 0.89)
- A Broad Study on the Transferability of Visual Representations With Contrastive Learning (ICCV 2021 · similarity 0.89)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (ICML 2025 · similarity 0.88)
- Divide and Contrast: Source-free Domain Adaptation via Adaptive Contrastive Learning (NeurIPS 2022 · similarity 0.88)
- Domain Adaptation with Invariant Representation Learning: What Transformations to Learn? (NeurIPS 2021 · similarity 0.88)
- SimProF: A Simple Probabilistic Framework for Unsupervised 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.
