Research map / Learning theory and algorithms / Kernels and regression theory
A kernel test for quasi-independence
Tamara Fernandez, Wenkai Xu, Marc Ditzhaus, Arthur Gretton
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 KQIC instead of WLR_SC | competing test · Monotonic Gaussian copula n=100 | rejection rate: 0.85 -> 0.93 | Improved |
| Increase sample size from n=100 to n=200 (ablation) | KQIC · Monotonic Gaussian copula rho=0.4 | rejection rate: 0.86 -> 1.00 | Improved |
| Increase sample size from n=100 to n=200 (ablation) | KQIC · Monotonic Gaussian copula rho=0.0 (null) | rejection rate: 0.06 -> 0.05 | No clear effect |
| Increase periodic frequency beta | KQIC and permutation tests · Periodic dependency model | power | Hurt |
| Use KQIC instead of permutation-based tests at high beta | KQIC · Periodic dependency, large beta | power | Improved |
| Increase censoring percentage | KQIC with Gaussian and IMQ kernels · Periodic dependency, sample size 800 | power: 85% | No clear effect |
| Use IMQ kernel instead of Gaussian kernel (ablation) | KQIC · High frequency dependency | power | No clear effect |
| Use KQIC instead of permutation tests | KQIC · Runtime | runtime | Improved |
| Use KQIC_Gauss instead of WLR_SC | competing test · Abortion Times combined | p-value: 0.511 -> 0.014 | Improved |
Closest papers by meaning
- Measuring Dependence Powerfully and Equitably (JMLR 2016 · similarity 0.84)
- A Distribution Free Conditional Independence Test with Applications to Causal Discovery (JMLR 2022 · similarity 0.84)
- PAIR-CI: Calibrated Conditional Independence Testing for Causal Discovery with Incomplete Data (NeurIPS 2026 · similarity 0.83)
- Kernelized Stein Discrepancy Tests of Goodness-of-fit for Time-to-Event Data (ICML 2020 · similarity 0.83)
- Independence Test for Linear Non-Gaussian Data and Applications in Causal Discovery (ICLR 2026 · similarity 0.83)
- Assessing the robustness of heterogeneous treatment effects in survival analysis under informative censoring (NeurIPS 2026 · similarity 0.82)
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.
