Machine Learning and Security

Machine Learning (ML) and Artificial Intelligence (AI) are becoming mainstream techniques, and they provide a great opportunity for defenders.


We are on the cusp of a Machine Learning and Artificial Intelligence revolution. ML and AI techniques have recently re-emerged as powerful tools in various business sectors such as Fraud Detection, Anomaly Detection, and Behavioral Analysis. Several companies and services are exploring these technologies and use them to solve specific security challenges successfully.

Despite the success of ML and AI, there are security risks associated with them, especially during the learning phase which can be vulnerable to threats originated by potential adversaries, with consequent impact on prediction results.

This Working Session will share common practices; what works today, and what is worth focusing on in the future.


  • What are the available machine learning platforms?
  • Are there any security vulnerabilities associated with these platforms?
  • How to securely feed data to ML and AI tools
  • How to make learning algorithms aware of malicious data?
  • Can AI be used to reduce false positive findings in security scanners?
  • How can we spread the message among developers and security communities?


  • Guidelines for secure usage of machine learning techniques.


The target audience for this Working Session is:

  • Security professionals
  • ML and AI researchers
  • Devops
  • SOC teams

Working materials

The content will have introduction about ML techniques, then we will try to resolve one exercise. after that we will go into discussion about the mentioned topics of this session.


Back to list of all Working Sessions and Tracks

Edit this page here