Octavian Suciu

PhD Student in Computer Science

University of Maryland, College Park

Bio

I am a PhD student in the Maryland Cybersecurity Center. My advisor is Prof. Tudor Dumitraş.

My research focuses on building robust and interpretable predictive systems for security. I am interested in the discovery of new cyberthreat indicators from unstructured data and the security of machine learning.

Interests

Security
Machine Learning

Education

MS in Computer Science, 2017

University of Maryland, College Park
BS in Computer Science, 2014

Technical University of Cluj-Napoca, Romania

Publications

Octavian Suciu, Scott E. Coull, Jeffrey Johns. Exploring Adversarial Examples in Malware Detection. ALEC’18 (AAAI 2018 Fall Symposium), 2018.

Venue

Octavian Suciu, Radu Mărginean, Yiğitcan Kaya, Hal Daumé III, Tudor Dumitraş. When Does Machine Learning FAIL? Generalized Transferability for Evasion and Poisoning Attacks. USENIX’18, 2018.

PDF Code Slides Technical Report Presentation

Ali Shafahi, W. Ronny Huang, Mahyar Najibi, Octavian Suciu, Christoph Studer, Tudor Dumitraş, Tom Goldstein. Poison Frogs! Targeted Clean-Label Poisoning Attacks on Neural Networks. arXiv, 2018.

Preprint PDF Code

Tudor Dumitraş, Yiğitcan Kaya, Radu Mărginean, Octavian Suciu. Too Big to FAIL: What You Need to Know Before Attacking a Machine Learning System. SPW’18, 2018.

Venue

Rock Stevens, Octavian Suciu, Andrew Ruef, Sanghyun Hong, Michael Hicks, Tudor Dumitraş. Summoning Demons: The Pursuit of Exploitable Bugs in Machine Learning. NIPS’16 Workshop on Reliable Machine Learning in the Wild, 2016.

Preprint PDF Slides Presentation Venue Media Coverage

Carl Sabottke, Octavian Suciu, Tudor Dumitraş. Vulnerability Disclosure in the Age of Social Media: Exploiting Twitter for Predicting Real-World Exploits. USENIX’15, 2015.

PDF Slides Presentation Blog Feature Set Explore Data

Contact

osuciu AT umiacs DOT umd DOT edu
@osuciu on Keybase.
3400 A.V. Williams Bldg, College Park, MD 20740, USA