Research map / Learning theory and algorithms / Kernels and regression theory
Hypothesis Transfer Learning via Transformation Functions
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 |
|---|---|---|---|
| Use an offset transformation instead of target-only KRR | Kernel Smoothing · kin-8fm → kin-8nh, n_ta = 40 | mean squared error: 0.063 ± 0.013 -> 0.052 ± 0.006 | Helped |
| Use an offset transformation instead of target-only KRR | Kernel Ridge Regression · kin-8fm → kin-8nh, n_ta = 80 | mean squared error: 0.050 ± 0.007 -> 0.048 ± 0.006 | Helped |
| Use an offset transformation instead of target-only KRR | Kernel Ridge Regression · kin-8fm → kin-8nh, n_ta = 320 | mean squared error: 0.040 ± 0.005 -> 0.041 ± 0.003 | Hurt |
| Use an offset transformation rather than a scale transformation | Transfer learning · kin-8fm → kin-8nh | mean squared error | Mixed |
| Use an offset transformation instead of target-only KRR | Kernel Smoothing · fMRI Stroop response-time prediction | R-squared (mean): 0.1041 -> 0.1190 | Helped |
| Use an offset transformation rather than a scale transformation | Transfer learning · fMRI Stroop response-time prediction | R-squared (mean): 0.0017 -> 0.1190 | Helped |
| Use an offset transformation rather than a scale transformation | Transfer learning · fMRI Stroop response-time prediction | R-squared standard deviation: 0.0632; 0.1104 | Mixed |
| Choose among transformation functions by cross-validation, including a no-transfer option | Transformation-function selection · Target-domain regression | excess risk | No clear effect |
| Use a transformation when it makes the auxiliary function smoother than the target regression function | Kernel Smoothing HTL · Target-domain regression | excess risk convergence rate | Helped |
| Use transfer learning when the auxiliary function has at least the target function's approximation regularity and source data is sufficiently large | Kernel Ridge Regression HTL · Target-domain regression | excess risk convergence rate | Helped |
Closest papers by meaning
- On Hypothesis Transfer Learning of Functional Linear Models (ICML 2024 · similarity 0.88)
- Wasserstein Transfer Learning (NeurIPS 2025 · similarity 0.87)
- Adversarially Robust Hypothesis Transfer Learning (ICML 2024 · similarity 0.87)
- Smoothness Adaptive Hypothesis Transfer Learning (ICML 2024 · similarity 0.87)
- Transfer Learning for Benign Overfitting in High-Dimensional Linear Regression (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 0.85)
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.