Research map / Label-efficient and robust learning / Domain adaptation and generalization
POEM: Polarization of Embeddings for Domain-Invariant Representations
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 POEM polarization to ERM | ERM / POEM · Five DG benchmarks (average) | accuracy: +1.5% | Helped |
| Add POEM to SWAD | SWAD · Five DG benchmarks (average) | accuracy: +0.9% | Helped |
| Add POEM to MIRO + SWAD | MIRO + SWAD · Five DG benchmarks (average) | accuracy: +0.6% | Helped |
| Add the disentangling loss L_s alone (ablation) | ERM-based POEM · Five DG benchmarks (average) | accuracy: +0.3% | Helped |
| Add the discrimination loss L_d alone (ablation) | ERM-based POEM · Five DG benchmarks (average) | accuracy: +0.5% | Helped |
| Combine L_s and L_d (ablation) | ERM-based POEM · Five DG benchmarks (average) | accuracy: +1.5% | Helped |
| Add POEM polarization to ERM | ERM / POEM · DomainNet | accuracy: 44.1 $\pm$ 0.0 -> 44.0 $\pm$ 0.0 | Hurt |
| Add POEM to SWAD | SWAD · OfficeHome | accuracy: 70.7 $\pm$ 0.1 -> 70.5 $\pm$ 0.1 | Hurt |
| Add POEM to MIRO + SWAD | MIRO + SWAD · PACS | accuracy: 87.7 $\pm$ 0.3 -> 88.5 $\pm$ 0.1 | Helped |
Closest papers by meaning
- Disentangled Prompt Representation for Domain Generalization (CVPR 2024 · similarity 0.86)
- Domain Generalization in CLIP via Learning with Diverse Text Prompts (CVPR 2025 · similarity 0.86)
- Multimodal Unsupervised Domain Generalization by Retrieving Across the Modality Gap (ICLR 2025 · similarity 0.85)
- Is Large-scale Pretraining the Secret to Good Domain Generalization? (ICLR 2025 · similarity 0.85)
- CLIPCEIL: Domain Generalization through CLIP via Channel rEfinement and Image-text aLignment (NeurIPS 2024 · similarity 0.85)
- Unknown Prompt the only Lacuna: Unveiling CLIP's Potential for Open Domain Generalization (CVPR 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.