Machine learning adversarial attacks are a ticking time bomb

By KDnuggets - 2021-03-19

Description

Software developers and cyber security experts have long fought the good fight against vulnerabilities in code to defend against hackers. A new, subtle approach to maliciously targeting machine learni ...

Summary

  • Software developers and cyber security experts have long fought the good fight against vulnerabilities in code to defend against hackers.
  • Security researchers are used to perusing code for vulnerabilities.
  • Since GitHub has no procedure to detect and block adversarial vulnerabilities, a malicious actor could easily use these kinds of tools to publish deep learning models that have hidden backdoors and exploit them after thousands of developers integrate them in their applications.
  • Bug bounties for adversarial vulnerabilities can also be a good measure to make sure the machine learning systems used by millions of users are robust.

 

Topics

  1. Machine_Learning (0.43)
  2. Backend (0.22)
  3. NLP (0.17)

Similar Articles

Applications of Machine Learning in FinTech

By Medium - 2016-04-05

Machine learning is a type of artificial intelligence that provides computers with the ability to learn without being explicitly programmed. The science behind machine learning is interesting and…