Poisoning Attacks against Machine Learning: Can Machine Learning be Trustworthy?

Published

October 24, 2022

Author(s)

Alina Oprea, Anoop Singhal, Apostol Vassilev

Abstract

Many practical applications benefit from Machine Learning (ML) and Artificial Intelligence (AI) technologies, but their security needs to be studied in more depth before the methods and algorithms are actually deployed in critical settings. In this article, we discuss the risk of poisoning attacks when training machine learning models and discuss challenges for defending against this threat.

Citation

Computer (IEEE Computer)

Volume

55

Issue

11

Pub Type

Journals

Download Paper

https://doi.org/10.1109/MC.2022.3190787

Local Download

Keywords

artificial intelligence technologies, machine learning trustworthiness, poisoning attacks, security

Citation

Oprea, A. , Singhal, A. and Vassilev, A. (2022), Poisoning Attacks against Machine Learning: Can Machine Learning be Trustworthy?, Computer (IEEE Computer), [online], https://doi.org/10.1109/MC.2022.3190787, https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=934932 (Accessed April 26, 2024)