Understanding Double Descent Requires A Fine-Grained Bias-Variance Decomposition

Authors: Ben Adlam, Jeffrey Pennington

NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Figure 2: Comparison of (a) ensembles and (b) bagging. Solid lines are theoretical predictions and dots are simulation results.
Researcher Affiliation Industry Ben Adlam Jeffrey Pennington* Google Brain {adlam, jpennin}@google.com
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statements about open-sourcing code or links to a code repository for the methodology described.
Open Datasets No The paper describes a synthetic data generation process for its analysis: 'We consider the task of learning an unknown function from m independent samples (xi, yi) 2 Rn0 R, i = 1, . . . , m, where the datapoints are standard Gaussian, xi N(0, In0), and the labels are generated by a linear function parameterized by β 2 Rn0, whose entries are drawn independently from N(0, 1).'
Dataset Splits No The paper does not provide specific dataset split information (percentages, sample counts, or predefined splits) for training, validation, or testing.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory, or cloud instance types) used for running its experiments or simulations.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes In (a,b) we set γ = 10 6, n0 = 213, m = 214, σ = tanh, and SNR = 5.