Machine Learning and Security

Back to list of all Outcomes

Original Working Session content: Machine Learning and Security


Synopsis and Takeaways

  • Create common datasets with the purpose of testing and validating the security of machine learning algorithms
    • We can use data output of Mod Security, WebGoat and others to create the datasets
    • These datasets should be shared
    • Anonymized dataset
    • Common dataset for testing
  • Create guidance page to include ML security definitions, latest reports, and links to the available tools and datasets
    • Find good materials and resources
  • Use ML techniques in the current tools provided by OWASP (e.g., use ML to reduce false positives in ZAP scanning output)

  • Create a working group to work on tools and guidance of:
    • How to check if a dataset is noise-free (not compromised)
    • Review of algorithms implementations


Back to list of all Outcomes

Edit this page here