Beyond Heuristics: Learning to Classify Vulnerabilities and Predict Exploits

   Abstract