During my PhD at Columbia, my research focused on the development of automatic security tools for network administrators. Network admission and access control is typically based on manually pre-determined rules and roles. However, as networks evolve into dynamic environments, new policies and roles need to be created in real time. Unfortunately, performing these tasks manually becomes a daunting task for network admins. My research designed and tested tools that automatically create and update such policies from behavioral traces. Specifically, I used data mining and machine learning techniques to reveal behavioral fingerprints from large network traffic datasets.