Hypothesis Set Stability and Generalization

Authors: Dylan J. Foster, Spencer Greenberg, Satyen Kale, Haipeng Luo, Mehryar Mohri, Karthik Sridharan

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We present a study of generalization for data-dependent hypothesis sets. We give a general learning guarantee for data-dependent hypothesis sets based on a notion of transductive Rademacher complexity. Our main result is a generalization bound for data-dependent hypothesis sets expressed in terms of a notion of hypothesis set stability and a notion of Rademacher complexity for data-dependent hypothesis sets that we introduce. This bound admits as special cases both standard Rademacher complexity bounds and algorithm-dependent uniform stability bounds. We also illustrate the use of these learning bounds in the analysis of several scenarios.
Researcher Affiliation Collaboration Dylan J. Foster Massachusetts Institute of Technology dylanf@mit.edu Spencer Greenberg Spark Wave admin@sparkwave.tech Satyen Kale Google Research satyen@satyenkale.com Haipeng Luo University of Southern California haipengl@usc.edu Mehryar Mohri Google Research and Courant Institute mohri@google.com Karthik Sridharan Cornell University sridharan@cs.cornell.edu
Pseudocode No The paper describes theoretical concepts and bounds but does not include any pseudocode or algorithm blocks.
Open Source Code No The paper does not include any statements about releasing source code or links to a code repository for the methodology described.
Open Datasets No The paper is theoretical and does not use or describe any specific datasets for empirical evaluation, thus no public dataset access information is provided.
Dataset Splits No The paper is theoretical and does not involve empirical experiments, thus no dataset split information for training, validation, or testing is provided.
Hardware Specification No The paper is theoretical and does not conduct experiments, thus no hardware specifications are mentioned.
Software Dependencies No The paper is theoretical and does not conduct experiments, thus no specific software dependencies with version numbers are listed.
Experiment Setup No The paper is theoretical and does not conduct experiments, thus no details on experimental setup or hyperparameters are provided.