Research map / Label-efficient and robust learning / Domain adaptation and generalization
Multivariate Regression on the Grassmannian for Predicting Novel Domains
Check 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 |
|---|---|---|---|
| Predict the target subspace with direct kernel regression and use GFK instead of no domain adaptation | VGG-full classifier · Comprehensive Cars Part-I, 30-domain hold-one-out | accuracy: 54.06 ± 12.09 -> 58.12 ± 9.69 | Helped |
| Predict distances to observed subspaces, infer the target subspace, and use GFK instead of no domain adaptation | VGG-full classifier · Comprehensive Cars Part-I, 30-domain hold-one-out | accuracy: 54.06 ± 12.09 -> 58.21 ± 9.56 | Helped |
| Predict the target subspace from domain metadata with the direct method rather than use its ground-truth subspace | VGG-full classifier with GFK · Comprehensive Cars Part-I, 30-domain hold-one-out | accuracy: 58.23 ± 9.68 -> 58.12 ± 9.69 | No clear effect |
| Predict the target subspace from domain metadata with the indirect method rather than use its ground-truth subspace | VGG-full classifier with GFK · Comprehensive Cars Part-I, 30-domain hold-one-out | accuracy: 58.23 ± 9.68 -> 58.21 ± 9.56 | No clear effect |
| Predict future subspaces using the direct or indirect method rather than Rolling UDA | Surveillance scene classifier · Surveillance scene classification, domains τ1–τ8 | precision | Helped |
| Use indirect rather than direct regression to predict subspaces | Surveillance scene classifier · Surveillance scene classification, domains τ1–τ8 | precision | Helped |
| Use indirect rather than direct regression to predict far-future subspaces | Surveillance scene classifier · Surveillance scene classification, four future domains after τ4 | precision | Helped |
| Predict domains from metadata rather than use the No-DA classifier directly | Surveillance scene classifier · Surveillance scene classification | precision | Helped |
| Observe more domains to train the Grassmannian regression model | Direct and Indirect predictive domain adaptation · Comprehensive Cars Part-I, two representative source-to-target pairs | accuracy | Helped |
Closest papers by meaning
- Learning an Invariant Hilbert Space for Domain Adaptation (CVPR 2017 · similarity 0.90)
- LEARNING FACTORIZED REPRESENTATIONS FOR OPEN-SET DOMAIN ADAPTATION (ICLR 2019 · similarity 0.89)
- Visual Representation Learning over Latent Domains (ICLR 2022 · similarity 0.89)
- Adaptive Methods for Real-World Domain Generalization (CVPR 2021 · similarity 0.89)
- Representation Subspace Distance for Domain Adaptation Regression (ICML 2021 · similarity 0.89)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (ICML 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.