Generalization of ERM in Stochastic Convex Optimization: The Dimension Strikes Back

Authors: Vitaly Feldman

NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this work we substantially strengthen the lower bound in [18] proving that a linear dependence on the dimension d is necessary for ERM (and, consequently, uniform convergence). We then extend the lower bound to all ℓp/ℓq setups and examine several related questions. Finally, we examine a more general setting of bounded-range SCO (that is |f(x)| 1 for all x K).
Researcher Affiliation Industry Vitaly Feldman IBM Research Almaden
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code No The paper is theoretical and does not mention releasing source code for its methodology.
Open Datasets No This is a theoretical paper and does not involve training on datasets.
Dataset Splits No This is a theoretical paper and does not involve dataset splits for validation.
Hardware Specification No The paper does not describe experiments and therefore provides no hardware specifications.
Software Dependencies No The paper does not describe experiments and therefore provides no software dependencies.
Experiment Setup No The paper does not describe experiments and therefore provides no experimental setup details.