On the Robustness of Kernel Ridge Regression Using the Cauchy Loss Function
Hongwei Wen, University of Twente, Annika Betken, University of Twente, Wouter Koolen, University of Twente · 2025
Read On the Robustness of Kernel Ridge Regression Using the Cauchy Loss Function by Hongwei Wen, University of Twente, Annika Betken, University of Twente, Wouter Koolen, University of Twente (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-20120.txt.
Description
Real data misbehave: outliers and wild noise break regression methods tuned for tidy errors. Hongwei Wen, Annika Betken, and Wouter Koolen study a kernel method armoured with the Cauchy loss, a penalty that refuses to chase extreme points. Earlier theory demanded noise with a finite average size, an assumption Cauchy and Pareto noise themselves violate. So the authors built a broader framework covering noise with any finite moment, even an infinite mean.
Inside it, they prove the method's error bounds approach the best possible rate, and that shrinking one tuning knob, the loss's scale, tightens performance in a controlled trade-off. Experiments on synthetic and real data under varied corruption confirmed the theory: the Cauchy armour holds where gentler losses dent.
Catalog record
| CREATOR | Hongwei Wen, University of Twente, Annika Betken, University of Twente, Wouter Koolen, University of Twente |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 116 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-20120