Back to search

Georgia Institute of Technology

Detecting and Leveraging Changes in Temporal Data

Abstract

dc:description.abstract

The ability to identify distributional shifts in temporal data is a common theme for both change-point detection and various machine learning tasks. This thesis explores the interplay between machine learning and change-point detection, and uses this interplay to devise new methods that utilize change-point detection for improving machine learning and in the reverse case, utilizes machine-learning to devise improved change point detection methods. For improving change-point detection, we explore how supervision can help learn distance metrics for designing robust change-point detection methods. For improving machine learning, we explore how change-points can provide weak supervision for supervised machine learning tasks. We also investigate how machine learning models can adapt to distributional changes by learning to selectively mask time series channels with significant shifts.

Degree

thesis:*
Level thesis:degree_level
Doctoral
Department dc:contributor.department
Electrical and Computer Engineering
Grantor dc:publisher
Georgia Institute of Technology
Year dc:date.issued
2024

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Ahad, Nauman
Advisor dc:contributor.advisor
  • Davenport, Mark A.
Committee members dc:contributor.committeemember
  • Xie, Yao
  • Romberg, Justin
  • Dyer, Eva
  • Inan, Omer
  • Sonenblum, Sharon

Subjects

dc:subject × 2

Rights

Language dc:language.iso
en_US

Identifiers

dc:identifier.*
Handle dc:identifier.uri
https://hdl.handle.net/1853/75667

Chain of custody

source
Harvested from
Georgia Tech
Base URL
repository.gatech.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Ahad, Nauman. Detecting and Leveraging Changes in Temporal Data. Doctoral thesis, Georgia Institute of Technology, 2024. https://hdl.handle.net/1853/75667