Tien Phan – “User identification via neural network based language model”

The idea is simple:

• If you read a lots of emails/messages from a person, you might be able to recognize that person’s emails/messages.
• So you want to train computers to do the same thing.

The application is open and depends on our creativity. Including, but not limited to:

• User/Group of users identification.
• Actively detection if the same user is using the account (which is not possible with username+password or two/three factors)
• Detect if an account is compromised and used to spread spam/phishing.

The content of the talk:

• Ideas/motivations
• Methodology
  • Approach
  • Dataset
  • Machine learning techniques/algorithms
• Lots of visualizations
• Some demos on the extracted machine learning features
• Results
• Conclusion

Community Presentations:

Julien Savoie – Title: Defeating IPv6 privacy extensions

Abstract: This will be a brief follow up to last months talk on general IPv6 security. Given that privacy addresses are selected at random, and not derived from a MAC address, tracking abuse can prove difficult.  While in corporate environments, it’s feasible to disable such privacy addresses, in BYOD or public access networks this is not the case. What follows will be a number of solutions for tracking these random addresses back to machines and ultimately users.

Moussa Noun – Title: How to Communicate Cyber Security to the Business

Abstract: Over the course of my years in pen-testing as a social engineer, I’ve learned a few tricks that I’ve ported over to my career as a cyber security professional in a business.