Generalization in Kernel Regression Under Realistic Assumptions

arXiv:2312.15995v2 Announce Type: replace
Abstract: It is by now well-established that modern over-parameterized models seem to elude the bias-variance tradeoff and generalize well despite overfitting noise. Many recent works attempt to analyze this phenomenon in the relatively tractable setting of kernel regression. However, as we argue in detail, most past works on this topic either make unrealistic assumptions, or focus on a narrow problem setup. This work aims to provide a unified theory to upper bound the excess risk of …

abstract analyze arxiv assumptions bias bias-variance cs.ai cs.lg kernel modern noise overfitting regression stat.ml tractable type variance