Online Safety Assurance for Learning-Augmented Systems

Noga H. Rotman, Michael Schapira, Aviv Tamar

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

8 Scopus citations

Abstract

Recently, deep learning has been successfully applied to a variety of networking problems. A fundamental challenge is that when the operational environment for a learning-augmented system differs from its training environment, such systems often make badly informed decisions, leading to bad performance. We argue that safely deploying learning-driven systems requires being able to determine, in real-time, whether system behavior is coherent, for the purpose of defaulting to a reasonable heuristic when this is not so. We term this the online safety assurance problem (OSAP). We present three approaches to quantifying decision uncertainty that differ in terms of the signal used to infer uncertainty. We illustrate the usefulness of online safety assurance in the context of the proposed deep reinforcement learning (RL) approach to video streaming. While deep RL for video streaming bests other approaches when the operational and training environments match, it is dominated by simple heuristics when the two differ. Our preliminary findings suggest that transitioning to a default policy when decision uncertainty is detected is key to enjoying the performance benefits afforded by leveraging ML without compromising on safety.

Original languageAmerican English
Title of host publicationHotNets 2020 - Proceedings of the 19th ACM Workshop on Hot Topics in Networks
PublisherAssociation for Computing Machinery, Inc
Pages88-95
Number of pages8
ISBN (Electronic)9781450381451
DOIs
StatePublished - 4 Nov 2020
Externally publishedYes
Event19th ACM Workshop on Hot Topics in Networks, HotNets 2020 - Virtual, Online, United States
Duration: 4 Nov 20206 Nov 2020

Publication series

NameHotNets 2020 - Proceedings of the 19th ACM Workshop on Hot Topics in Networks

Conference

Conference19th ACM Workshop on Hot Topics in Networks, HotNets 2020
Country/TerritoryUnited States
CityVirtual, Online
Period4/11/206/11/20

Bibliographical note

Publisher Copyright:
© 2020 ACM.

Keywords

  • network protocol design
  • reinforcement learning
  • sequential decision making
  • video streaming

Fingerprint

Dive into the research topics of 'Online Safety Assurance for Learning-Augmented Systems'. Together they form a unique fingerprint.

Cite this