On the S-instability and degeneracy of discrete deep learning models

Date
2019-11-05
Authors
Kaplan, Andee
Vardeman, Stephen
Nordman, Daniel
Vardeman, Stephen
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Statistics
Organizational Unit
Journal Issue
Series
Department
StatisticsIndustrial and Manufacturing Systems Engineering
Abstract

A probability model exhibits instability if small changes in a data outcome result in large and, often unanticipated, changes in probability. This instability is a property of the probability model, given by a distributional form and a given configuration of parameters. For correlated data structures found in several application areas, there is increasing interest in identifying such sensitivity in model probability structure. We consider the problem of quantifying instability for general probability models defined on sequences of observations, where each sequence of length N has a finite number of possible values that can be taken at each point. A sequence of probability models, indexed by N⁠, and an associated parameter sequence result to accommodate data of expanding dimension. Model instability is formally shown to occur when a certain log probability ratio under such models grows faster than N⁠. In this case, a one component change in the data sequence can shift probability by orders of magnitude. Also, as instability becomes more extreme, the resulting probability models are shown to tend to degeneracy, placing all their probability on potentially small portions of the sample space. These results on instability apply to large classes of models commonly used in random graphs, network analysis and machine learning contexts.

Comments

This is a pre-copyedited, author-produced version of an article accepted for publication in Information and Inference: A Journal of the IMA following peer review. The version of record: Kaplan, Andee, Daniel J. Nordman, and Stephen B. Vardeman. "On the S-instability and degeneracy of discrete deep learning models," Information and Inference: A Journal of the IMA is available online at DOI: 10.1093/imaiai/iaz022. Posted with permission.

Description
Keywords
Citation
DOI
Collections