Research map / Label-efficient and robust learning / Domain adaptation and generalization
Multi-Source Domain Adaptation for Text Classification via DistanceNet-Bandits
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 a domain-distance loss to the source classification loss | DistanceNet · Single-source transfer, average of MR/Apparel/Baby domain pairs | accuracy: 70.08 -> 71.32 | Helped |
| Use a selected mixture of distance measures instead of one distance (ablation) | DistanceNet · Single-source transfer, average of MR/Apparel/Baby domain pairs | accuracy: 70.42 -> 71.32 | Helped |
| Use L2 distance as the added loss (ablation) | DistanceNet · Single-source transfer, average of MR/Apparel/Baby domain pairs | accuracy: 70.08 -> 70.37 | Helped |
| Use Cosine distance as the added loss (ablation) | DistanceNet · Single-source transfer, average of MR/Apparel/Baby domain pairs | accuracy: 70.08 -> 70.48 | Helped |
| Use MMD as the added loss (ablation) | DistanceNet · Single-source transfer, average of MR/Apparel/Baby domain pairs | accuracy: 70.08 -> 70.42 | Helped |
| Use Fisher distance as the added loss (ablation) | DistanceNet · Single-source transfer, average of MR/Apparel/Baby domain pairs | accuracy: 70.08 -> 70.10 | No clear effect |
| Dynamically select source domains using a bandit controller rather than round-robin selection | Multi-source DistanceNet · Average of five target domains | accuracy: 78.20 -> 79.30 | Helped |
| Dynamically select source domains using a bandit controller rather than round-robin selection | Multi-source DistanceNet · Five target domains: MR, Apparel, Baby, Books, and Camera | accuracy | Helped |
Closest papers by meaning
- Adversarial Training Based Multi-Source Unsupervised Domain Adaptation for Sentiment Analysis (AAAI 2020 · similarity 0.88)
- Domain Aggregation Networks for Multi-Source Domain Adaptation (ICML 2020 · similarity 0.87)
- Adversarial Multiple Source Domain Adaptation (NeurIPS 2018 · similarity 0.87)
- Bridging Theory and Algorithm for Domain Adaptation (ICML 2019 · similarity 0.87)
- Towards Domain Adaptive Neural Contextual Bandits (ICLR 2025 · similarity 0.84)
- Theoretical Performance Guarantees for Partial Domain Adaptation via Partial Optimal Transport (ICML 2025 · 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.