Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Transfer Learning with Uncertainty Quantification: Random Effect Calibration of Source to Target (RECaST)
Jimmy Hickey, Jonathan P. Williams, Emily C. Hector
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 |
|---|---|---|---|
| use RECaST LM instead of target-only DNN | DNN · Continuous, n_T=20, sigma^2=0 | RMSE: 20.5 -> 0.535 | Improved |
| use RECaST LM instead of target-only DNN | DNN · Continuous, n_T=250, sigma^2=4 | RMSE: 3.67 -> 14.3 | Hurt |
| use RECaST LM instead of Unfreeze DNN | Unfreeze DNN · Continuous, n_T=20, sigma^2=0 | RMSE: 2.49 -> 0.535 | Improved |
| use RECaST LM instead of Unfreeze DNN | Unfreeze DNN · Continuous, n_T=250, sigma^2=0.25 | RMSE: 2.76 -> 3.6 | Hurt |
| use RECaST DNN instead of target-only DNN | DNN · Binary, n_T=20, sigma^2=0 | AUC: 60.2 -> 97.3 | Improved |
| use RECaST DNN instead of target-only DNN | DNN · Binary, n_T=250, sigma^2=4 | AUC: 94.5 -> 89.4 | Hurt |
| use RECaST DNN instead of Wiens regularized logistic regression | Wiens · Binary, n_T=250, sigma^2=0 | AUC: 80.4 -> 98.3 | Improved |
| use RECaST DNN instead of Unfreeze DNN | Unfreeze DNN · Binary, n_T=20, sigma^2=4 | AUC: 77.1 -> 88.2 | Improved |
| use RECaST DNN instead of target-only DNN for prediction sets | DNN · Binary coverage, n_T=20, sigma^2=0 | coverage: 67.2 -> 95.6 | Improved |
| increase source-target dissimilarity sigma^2 from 0 to 4 | RECaST DNN · Binary, n_T=250 | AUC: 98.3 -> 89.4 | Hurt |
| reduce target sample size from 250 to 20 | RECaST LM · Continuous, sigma^2=0 | RMSE: 0.517 -> 0.535 | No clear effect |
| reduce target sample size from 250 to 20 | target-only DNN · Continuous, sigma^2=0 | RMSE: 2.8 -> 20.5 | Hurt |
Closest papers by meaning
- Wasserstein Transfer Learning (NeurIPS 2025 · similarity 0.86)
- Transfer Learning of Linear Regression with Multiple Pretrained Models: Benefiting from More Pretrained Models via Overparameterization Debiasing (NeurIPS 2026 · similarity 0.86)
- Transfer Learning for Benign Overfitting in High-Dimensional Linear Regression (NeurIPS 2025 · similarity 0.86)
- Transfer Learning Meets Functional Linear Regression: No Negative Transfer Under Posterior Drift (AAAI 2025 · similarity 0.85)
- A High-Dimensional Statistical Method for Optimizing Transfer Quantities in Multi-Source Transfer Learning (NeurIPS 2025 · similarity 0.85)
- Transfer Learning Through Conditional Quantile Matching (NeurIPS 2026 · 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.
