Target alignment in truncated kernel ridge regression

Authors: Arash Amini, Richard Baumgartner, Dai Feng

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We provide experiments verifying the multiple-descent and non-monotonic behavior of the regularization curves as well as the improved rate of Theorem 2 (Section 4.2). We present various simulation results to demonstrate the multiple-descent and phase transition behavior of the regularization curves, and corroborate the theoretical results.
Researcher Affiliation Collaboration Arash A. Amini1, Richard Baumgartner2, Dai Feng3 1University of California, Los Angeles 2Merck & Co., Inc., Rahway, New Jersey, USA 3Data and Statistical Sciences, Abb Vie Inc.
Pseudocode No The paper contains mathematical derivations and proofs, but no structured pseudocode or algorithm blocks were found.
Open Source Code Yes The code for reproducing the simulations is available at [3].
Open Datasets No The paper describes generating synthetic data for simulations ("200 samples generated from a uniform distribution on [0, 1]d") rather than using a publicly available dataset with specific access information.
Dataset Splits No The paper describes generating synthetic data for simulations but does not specify any explicit training, validation, or test dataset splits.
Hardware Specification No Our simulation were done on a regular laptop.
Software Dependencies No No specific software dependencies with version numbers (e.g., programming languages, libraries, or frameworks) were mentioned in the paper.
Experiment Setup No The paper describes parameters for its simulations such as "Gaussian kernel e x x /2h2 in d = 4 dimensions with bandwidth h = p d/2" and fixing λ for regularization curves. It also notes how random entries for ξ are generated. However, it does not explicitly list hyperparameters or system-level training settings in the typical sense for a machine learning model.