Research map / Label-efficient and robust learning / Domain adaptation and generalization
Budget-Aware Adapters for Multi-Domain Learning
Rodrigo Berriel, Stephane Lathuillere, Moin Nabi, Tassilo Klein, Thiago Oliveira-Santos, Nicu Sebe, Elisa Ricci
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 |
|---|---|---|---|
| reduce budget beta from 1.00 to 0.25 in BA2 (ablation) | BA2 · Visual Decathlon | S-Score: 3199 -> 2538 | Hurt |
| reduce budget beta from 1.00 to 0.25 in BA2 (ablation) | BA2 · Visual Decathlon | mean accuracy: 75.7 -> 72.6 | Hurt |
| reduce budget beta from 1.00 to 0.25 in BA2 (ablation) | BA2 · Visual Decathlon | S_O: 4952 -> 7809 | Improved |
| reduce budget beta from 1.00 to 0.25 in BA2 (ablation) | BA2 · Visual Decathlon | relative FLOP: 0.646 -> 0.325 | Improved |
| reduce budget beta from 1.00 to 0.25 in BA2 (ablation) | BA2 · ImageNet-to-Sketch | S-Score: 1265 -> 755 | Hurt |
| reduce budget beta from 1.00 to 0.25 in BA2 (ablation) | BA2 · ImageNet-to-Sketch | S_O: 1807 -> 2013 | Improved |
| reduce budget beta from 1.00 to 0.25 in BA2 (DenseNet-121) (ablation) | BA2 · ImageNet-to-Sketch (DenseNet-121) | S-Score: 1440 -> 840 | Hurt |
| reduce budget beta from 0.75 to 0.50 in BA2 (ablation) | BA2 · ImageNet-to-Sketch | S-Score: 1006 -> 1012 | No clear effect |
| use SpotTune-style per-domain full fine-tuning instead of BA2 beta=1.00 | SpotTune · Visual Decathlon | S-Score: 3199 -> 3612 | Improved |
| use independent switches per budget (BA2) instead of nested filters (Slimmable Networks) (ablation) | BA2 · CIFAR-100 | accuracy | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- SpotTune — SpotTune: Transfer Learning Through Adaptive Fine-Tuning (CVPR 2019) · better on 0 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Domain Adaptive Multibranch Networks (ICLR 2020 · similarity 0.87)
- MultiDIAL: Domain Alignment Layers for (Multisource) Unsupervised Domain Adaptation (TPAMI 2020 · similarity 0.86)
- Moment Matching for Multi-Source Domain Adaptation (ICCV 2019 · similarity 0.86)
- AdaGraph: Unifying Predictive and Continuous Domain Adaptation Through Graphs (CVPR 2019 · similarity 0.85)
- Multi-Source Distilling Domain Adaptation (AAAI 2020 · similarity 0.85)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (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.
