Research map / Label-efficient and robust learning / Domain adaptation and generalization
Domain-Indexing Variational Bayes: Interpretable Domain Index 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 |
|---|---|---|---|
| Infer domain indices with VDI (adversarial variational Bayes) instead of training without domain indices | domain adaptation on Circle · Circle | Accuracy (%): 55.5 -> 94.3 | Helped |
| Infer domain indices with VDI instead of baselines | DG-15 and DG-60 · DG-15 / DG-60 | Accuracy (%): DG-15 best baseline 79.9 -> 94.7; DG-60 best baseline 82.1 -> 95.9 | Helped |
| Infer domain indices with VDI | TPT-48 W(6)->E(42) regression · TPT-48 W(6)->E(42) | MSE (all 42 domains): 4.576 (Source-Only) -> 2.496 | Helped |
| Infer domain indices with VDI | TPT-48 N(24)->S(24) regression · TPT-48 N(24)->S(24) | MSE (all 24 domains): 4.560 (Source-Only) -> 3.160 | Helped |
| Most standard DA baselines (DANN, ADDA, MDD, SENTRY) using only domain identities | TPT-48 regression · TPT-48 both tasks | MSE (all domains): e.g. ADDA 6.528 vs Source-Only 4.560 on N->S | Hurt |
| Infer domain indices with VDI | CompCars 4-way classification · CompCars | Accuracy (%): 39.1 (Source-Only) -> 43.9; best baseline ADDA 42.8 | Helped |
| Domain adversarial training with the discriminator on z (adding DANN-style adversary) in the standard DA baselines | DANN on DG-15 · DG-15 | Accuracy (%): DANN 43.4 vs Source-Only 39.7 on DG-15; DANN 66.3 on DG-60 vs 55.0 Source-Only, but 53.4 vs 55.5 on Circle | Mixed |
Closest papers by meaning
- Bayesian Domain Adaptation with Gaussian Mixture Domain-Indexing (NeurIPS 2024 · similarity 0.89)
- Variational Model Perturbation for Source-Free Domain Adaptation (NeurIPS 2022 · similarity 0.88)
- Bridging Theory and Algorithm for Domain Adaptation (ICML 2019 · similarity 0.87)
- Continuously Indexed Domain Adaptation (ICML 2020 · similarity 0.87)
- Revisiting the Domain Shift and Sample Uncertainty in Multi-source Active Domain Transfer (CVPR 2024 · similarity 0.84)
- On $f$-Divergence Principled Domain Adaptation: An Improved Framework (NeurIPS 2024 · similarity 0.84)
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.