Reading group

Meeting format

• Frequency: every 2 weeks.
• Duration: 60 minutes.
• Participants: ML team, colleagues from Positive Technologies, and external participants.

Speaker role

Each session is led by a volunteer (write in the channel chat if you want to present).

Speaker responsibilities

• Choose an article (from proposed options or your own) no later than 7 days before the session.
• Prepare supporting materials: slides, code examples, and discussion prompts.
• Moderate the discussion.

Topic selection guidelines

Core directions

• ML/DL in cybersecurity (threat analysis, anti-fraud, NLP for log analysis).
• New algorithms (transformers, explainable AI, adversarial attacks).
• Adjacent domains: DevSecOps, DE, cryptography, ethical hacking.

Materials criteria

• Relevance: typically no older than 2-3 years unless it is a classic.
• Practical value for AI or Cybersecurity industries.
• Clarity of presentation: avoid overly theoretical papers with no examples.