AI Paper Atlas

Research map / Learning theory and algorithms / Kernels and regression theory

Streaming kernel regression with provably adaptive mean, variance, and regularization

Audrey Durand, Odalric-Ambrym Maillard, Joelle Pineau

JMLR 2018 · Learning theory and algorithms › Kernels and regression theory · approach: Kernel ridge regression analysis · 29 citations (OpenAlex)

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 changedWhereMeasuredEffect
tune regularization lambda to sigma^2/C^2 instead of sigma^2 in the concentration bound (ablation)kernel confidence bound (Theorem) · Synthetic 1D kernel regression (Gaussian kernel, sigma=0.1)confidence envelope widthImproved
use adaptive lambda = sigma_{+,t-1}^2/C^2 instead of fixed lambda = sigma_+^2/C^2 (ablation)noise upper-bound estimate · Synthetic 1D kernel regressionvariance estimate tightnessImproved
use fully adaptive regularization with estimated noise instead of fixed lambda with sigma_+ = 1 (ablation)kernel confidence interval · Synthetic 1D kernel regression (delta=0.1)confidence interval widthImproved
use adaptive regularization instead of fixed lambda = sigma_+^2/C^2 (ablation)Kernel UCB and Kernel TS · Kernelized bandit, 100 arms, sigma=0.1cumulative regretImproved
replace Kernel TS with Kernel UCBkernelized bandits · Kernelized bandit, 100 armscumulative regretImproved
use the new concentration bound instead of the earlier Lemma bound in Kernel TS oracleKernel TS · Kernelized bandit, 100 armscumulative regretImproved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy