Machine Learning for Computer Security

Special Issue in Journal of Machine Learning Research

Published Papers:

Call For Papers

As computers have become more ubiquitous and connected, their security has become a major concern. Of interest to this special issue is research that demonstrates how machine learning (or data mining) techniques can be used to improve computer security. This includes efforts directed at improving security of networks, hosts, and individual applications or computer programs. Research can have many goals including, but not limited to, authenticating users, characterizing the system being protected, detecting known or unknown vulnerabilities that could be exploited, using software repositories as training data to find software bugs, preventing attacks, detecting known and novel attacks when they occur, analyzing recently detected attacks, responding to attacks, predicting attacker actions and goals, performing forensic analysis of compromised systems, and analyzing activities seen in honey pots and network "telescopes" or "black holes."

Of special interest are studies that use machine learning techniques, carefully describe their approach, evaluate performance in a realistic environment, and compare performance to existing accepted approaches. Studies that use machine learning techniques or extend current techniques to address difficult security-related problems are of most interest.

It is expected that studies will have to address many classic machine learning issues including feature selection, feature construction, incremental/online learning, noise in the data, skewed data distributions, distributed learning, correlating multiple models, and efficient processing of large amounts of data.

Important Dates: (expired)

Submission Guidelines:

Guest Editors: Reviewers: