Research map / Label-efficient and robust learning / Domain adaptation and generalization
Transferable Calibration with Lower Bias and Variance in 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 |
|---|---|---|---|
| Apply TransCal instead of temperature scaling | CDAN · Office-Home (9 tasks; Avg) | ECE (%): 16.8 -> 13.5 | Helped |
| Apply TransCal instead of IID temperature scaling | MDD · Office-Home (6 tasks; Avg) | ECE (%): 17.5 -> 11.6 | Helped |
| Apply TransCal instead of CPCS | MCD · Office-Home (6 tasks; Avg) | ECE (%): 18.4 -> 12.0 | Helped |
| Apply TransCal instead of CPCS | MDD · ImageNet-Sketch (ImageNet→Sketch) | ECE (%): 14.2 -> 8.1 | Helped |
| Apply TransCal instead of CPCS | MCD · VisDA (Simulation→Real) | ECE (%): 22.9 -> 7.8 | Helped |
| Apply TransCal instead of temperature scaling | Multi-Domain Sentiment · 12 transfer tasks (Avg) | ECE (%): 9.2 -> 5.9 | Helped |
| Add learnable bias-reduction term (ablation) | TransCal with CDAN · Office-Home (9 tasks; Avg) | ECE (%): 14.5 -> 13.5 | Helped |
| Add variance-reduction term (ablation) | TransCal with CDAN · Office-Home (9 tasks; Avg) | ECE (%): 18.6 -> 13.5 | Helped |
| Add control variate (ablation) | TransCal with MDD · Office-Home A→C | ECE (%): 20.9±4.68 -> 13.5±3.51 | Helped |
| Use serial rather than parallel control variates (ablation) | TransCal with MDD · Office-Home A→C | ECE (%): 13.6±4.43 -> 13.5±3.51 | No clear effect |
| Use serial rather than parallel control variates (ablation) | TransCal with MDD · Office-Home A→P | ECE (%): 10.6±1.46 -> 11.4±0.81 | Mixed |
Closest papers by meaning
- Pseudo-Calibration: Improving Predictive Uncertainty Estimation in Unsupervised Domain Adaptation (ICML 2024 · similarity 0.90)
- Domain Adaptation with Conditional Transferable Components (ICML 2016 · similarity 0.89)
- Revisiting the Domain Shift and Sample Uncertainty in Multi-source Active Domain Transfer (CVPR 2024 · similarity 0.89)
- Dirichlet-based Uncertainty Calibration for Active Domain Adaptation (ICLR 2023 · similarity 0.89)
- Revisiting Source-Free Domain Adaptation: a New Perspective via Uncertainty Control (ICLR 2025 · similarity 0.88)
- Post-Hoc Uncertainty Calibration for Domain Drift Scenarios (CVPR 2021 · 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.