Odyssey: A Systems Approach to Machine Learning Security

reportActive / Technical Report | Accesssion Number: AD1157105 | Open PDF

Abstract:

This paper provides a systems approach to addressing attacks, consequences, and mitigations for systems using Machine Learning (ML). It explains each of these over the lifecycle of an ML technology, providing clear explanations of what to worry about, when to worry about it, and how to mitigate it while presuming little incoming knowledge of ML specifics. Our discussion of ML vulnerabilities, attacks, and mitigations utilizes the taxonomy developed in NISTIR 8269.

Author(s):
Winder, Ransom ; , ; Jubinski, Joseph ; Giannella, Chris ; Malachi, Jones
Author Organization(s):
MITRE CORP ANNAPOLIS JUNCTION MD
Funding Organization(s):
OFFICE OF THE SECRETARY OF DEFENSE WASHINGTON DC, WASHINGTON DC , DC
Document Type:
Technical Report/Technical Report
Publication Date:
2021 Apr 01
Pagination:
43

Security Markings

DOCUMENT & CONTEXTUAL SUMMARY

Distribution Code:
A - Approved For Public Release
Distribution Statement: Public Release

RECORD

Collection: TRECMS
Identifying Numbers
Contract Number(s):
W56KGU-18-D-0004
Project Number(s):
0720N05Z-0A
Subject Terms
Modernization Areas:
Autonomy
Communities of Interest:
Autonomy
Descriptor(s):
machine learning, systems approach, addressing, learning, taxonomy, vulnerability
Keyword(s):
AML(Adversarial Machine Learning), ML(machine learning), rl(Reinforcement Learning), ai(Artificial Intelligence), attack mitigation
Subject Categories:
Mathematical and Computer Sciences; Behavioral and Social Sciences
Creation Date:
2022 Jan 21
Update Date:
2022 Aug 30