On the Power of Over-parametrization in Neural Networks with Quadratic Activation

Authors: Simon Du, Jason Lee

ICML 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We provide new theoretical insights on why overparametrization is effective in learning neural networks.
Researcher Affiliation Academia 1Machine Learning Department, Carnegie Mellon University 2Department of Data Sciences and Operations, University of Southern California. Correspondence to: Simon S. Du <ssdu@cs.cmu.edu>.
Pseudocode No The paper does not contain any pseudocode or clearly labeled algorithm blocks.
Open Source Code No The paper does not provide any statements about releasing open-source code for the described methodology, nor does it include links to a code repository.
Open Datasets No The paper refers to 'n training data points' and discusses theoretical properties under assumptions like 'data is sampled from a regular distribution such as Gaussian.' It mentions 'arbitrary data set' and 'xi ~ N(0, I)' for theoretical analysis, but these are not actual, publicly accessible datasets used in empirical experiments.
Dataset Splits No The paper is theoretical and does not describe empirical experiments, thus no specific dataset splits (training, validation, test) are mentioned.
Hardware Specification No The paper is theoretical and does not describe empirical experiments, therefore no specific hardware specifications are mentioned.
Software Dependencies No The paper is theoretical and does not describe empirical experiments, therefore no specific software dependencies with version numbers are mentioned.
Experiment Setup No The paper is theoretical and does not describe empirical experiments, therefore no specific experimental setup details such as hyperparameters or training configurations are provided.