Mike Brennan.
mb553@drexel.edu.



I am a PhD candidate in the Computer Science Department Drexel University in Philadelphia. I am currently working with my advisor Rachel Greenstadt in the Privacy, Securlty and Automation Lab (PSAL). Currently we are working on stylometry (authorship recognition) and it's applications in artificial intelligence and security. We are specifically looking at methods of attacking stylometric systems, the potential threat authorship recognition poses to privacy and anonymity, and the mitigation of this threat through such attacks. We are also working on using AI to automatically replicate peer-reviewed trust metrics, such as the comment rating systems found on popular websites like Slashdot and YouTube. I have also worked with Frank Lee on experimental educational gaming.

I own a beautiful old house in West Philadelphia, I run a small diy record label, and enjoy spending smaller bits of free time biking around the city and larger bits of free time traveling a bit further from home.

Would you like an overview of my authorship recognition research? Check out this YouTube playlist containing my talk from 26C3 in December of 2009.


Publications.
"Learning to Extract Quality Discourse in Online Communities." Brennan/Wrazien/Greenstadt. WikiAI '10. [pdf]
"Using Machine Learning to Augment Collaborative Filtering of Community Discussions." Brennan/Wrazien/Greenstadt (Extended Abstract). AAMAS '10.
"Mixed-Initiative Security Agents." Greenstadt/Afroz/Brennan. AISec '09. [pdf]
"Practical Attacks Against Authorship Recognition Techniques." Brennan/Greenstadt. IAAI '09. [pdf]
"Museum Hunt: a Computerized Eye-Tracking Game." Brennan/Lee/Hurewitz/Boucher. International Meeting for Autism Research '09.
"An Adaptive Color Based Method for Robot Path Identification." Brennan. University of Delaware Senior Thesis '05.

Talks.
"Learning to Extract Quality Discourse in Online Communities." "Privacy & Stylometry: Practical Attacks Against Authorship Recognition Techniques." [slides: pdf, ppt]
Education & Awards.
Upsilon Pi Epsilon (Computer Science Honor Society) Member, 2008-.
Dean's Fellowship, Drexel University, 2007-.
B.S., Degree with Distinction, Computer Science, University of Delaware, 2005.

Miscellaneous.
"Writing style fingerprint easily fooled." by Colin Barras. New Scientist. (An overview of our Stylometry research)