Multiple Descent: Design Your Own Generalization Curve

Authors: Lin Chen, Yifei Min, Mikhail Belkin, Amin Karbasi

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical Our work proves that the expected risk of linear regression can manifest multiple descents when the number of features increases and sample size is fixed. This is carried out through an algorithmic construction of a feature-revealing process where the newly revealed feature follows either a Gaussian distribution or a Gaussian mixture distribution.
Researcher Affiliation Academia Lin Chen Simons Institute for the Theory of Computing University of California, Berkeley CA 94720 lin.chen@berkeley.edu Yifei Min Department of Statistics and Data Science Yale University CT 06511 yifei.min@yale.edu Mikhail Belkin Halıcıo glu Data Science Institute University of California, San Diego CA 92093 mbelkin@ucsd.edu Amin Karbasi School of Engineering and Applied Science Yale University CT 06511 amin.karbasi@yale.edu
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The ethics checklist states that code was not included for experimental results ('N/A'). No explicit statement or link for open-source code for the described methodology was found.
Open Datasets No The paper is theoretical and does not involve experiments on a specific dataset, thus no public dataset information for training is provided.
Dataset Splits No The paper is theoretical and does not involve empirical experiments, so no training/validation/test dataset splits are described.
Hardware Specification No The paper is theoretical and does not report experimental results, thus no specific hardware specifications are mentioned.
Software Dependencies No The paper is theoretical and does not report experimental results, thus no specific software dependencies with version numbers are mentioned.
Experiment Setup No The paper is theoretical and does not describe any empirical experiments, so no specific experimental setup details or hyperparameters are provided.