I am a Research Scientist at Facebook, focusing on deep learning uncertainty. My research aims at improving deep learning reliability in the presence of noisy labels for image classification and semantic segmentation. Prior to joining Facebook, I was staff research scientist and machine learning tech lead in Security and Privacy Research at Intel Labs. My research was in the intersection of deep learning, security and privacy with applications in threat intelligence, malware detection, fraud detection and computer vision. In 2016-2019, I was the principal investigator and research lead at the Intel Science & Technology Center on Adversary-Resilient Security Analytics, a collaboration between Intel and Georgia Tech.
My research has been published in pioneering journals and conferences including IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), Annals of Applied Statistics, Parallel Computing, International Conference on Machine Learning (ICML), IEEE Computer Vision Pattern Recognition (CVPR), KDD, IEEE Security and Privacy, ACM CCS, ACSAC, AAAI Conference on Artificial Intelligence, SPIE. My research on malware intelligence has been featured in top tech websites including Wall Street Journal, Engadget, Fortune, Tech Xplore, ZDNet, SecurityWeek, TechRadar and many more. I have given more than 60 technical presentations at conferences including BlackHat USA, BlueHat, DefCon, Embedded Vision Summit, Joint Statistical Meeting, AAAI, International Joint Conference on Artificial Intelligence and Spring Research Conference on Statistics and Industry Technology. I hold two granted patents with over 20 pending.
Deep learning uncertainty, noisy label robustness, adversarial machine learning, privacy-preserving machine learning