Machine Learning in Vulnerability Assessment

PDF

Authors
  1. Montuno, D.
Corporate Authors
Defence Research and Development Canada, Ottawa Research Centre, Ottawa ON (CAN);Solana Networks - Ottawa, ON (CAN)
Abstract
Machine Learning (ML) is increasingly being applied in vulnerability assessment and more generally in providing cyber security. We review ML applications in both those areas by commercial vendors. We also review recent results in adversarial learning; since ML requires training data to be effective, it is susceptible to adversarial attacks in which that data is poisoned to impair the ML’s functionality or allow attackers to bypass it. As a result of this adversarial nature of the problem, we conclude that the automated nature of ML-based solutions increases the need for accurate ground truth input data, and that more research is required to ensure the safety and effectiveness of these approaches with human-machine cooperation in mind.

Il y a un résumé en français ici.

Keywords
Cyber;Cyber Defence;Cyber Security;Machine Learning
Report Number
DRDC-RDDC-2019-C067 — Contract Report
Date of publication
01 Apr 2019
Number of Pages
21
DSTKIM No
CA049336
CANDIS No
810098
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: