Deep Reinforcement One-Shot Learning for Change Point Detection

Anton Puzanov, Kobi Cohen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

We consider the problem of detecting a change in a time series quickly and reliably, where only a few training instances are available. Examples include identifying changes in network traffic due to zero-day attacks, and computer vision applications where changes in series of images that represent significant events needed to be detected. These are known as cases of one-shot learning. We develop a novel Deep Reinforcement One-shot Learning (DeROL) framework to address this challenge. The basic idea of the DeROL algorithm is to train a deep-Q network to obtain a policy which is oblivious to the unseen classes in the testing data. Then, in real-time, DeROL maps the current state of the one-shot learning process to operational actions based on the trained deep-Q network, to maximize the objective function. We tested the algorithm using the OMNIGLOT dataset to demonstrate the efficiency of the DeROL framework.

Original languageEnglish
Title of host publication2018 56th Annual Allerton Conference on Communication, Control, and Computing, Allerton 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1047-1051
Number of pages5
ISBN (Electronic)9781538665961
DOIs
StatePublished - 5 Feb 2019
Event56th Annual Allerton Conference on Communication, Control, and Computing, Allerton 2018 - Monticello, United States
Duration: 2 Oct 20185 Oct 2018

Publication series

Name2018 56th Annual Allerton Conference on Communication, Control, and Computing, Allerton 2018

Conference

Conference56th Annual Allerton Conference on Communication, Control, and Computing, Allerton 2018
Country/TerritoryUnited States
CityMonticello
Period2/10/185/10/18

Fingerprint

Dive into the research topics of 'Deep Reinforcement One-Shot Learning for Change Point Detection'. Together they form a unique fingerprint.

Cite this