Mohsen Imani Email & Phone Number

Redwood City, California, United States

Redwood City, California, United States

Redwood City, California, United States

Foster City, Ca

Arlington, Texas

Working on improving the performance and security of anonymity systems. We have developed defenses to hinder Website Fingerprinting attacks in Tor network. Website fingerprinting attacks (WF) are machine learning problem that enable a local eavesdropper to determine which websites a user is visiting over an encrypted connection. Our defense could drop the accuracy rate of the state-of-the-art attack from 92% to 17%. We also designed approaches using different techniques, including Machine Learning algorithms, to improve the performance of Tor network. In our best approach we could improve the performance up to 20% compared to the state-of-the-art approach.

Arlington, Texas

CSE6388, Special Topics on advanced Information Security.CSE5388, Special Topics on Information Security.CSE5381, Information Security 2.CSE5380, Information Security 1.CSE1301, Computer Literacy.

United States

Working on security and safety of the autonomous and self-driving cars using ma- chine learning, detecting the crash and near crash events using machine learning techniques. The project focused on how the autonomous vehicles are resilient in the presence of an adversary. We implemented multiple Safety Event Detection (SED) systems. These SED systems used machine learning algorithms such as k-means, SVM, Neural Networks, and k-NN. The goal was to investigate how resilient the implemented systems are in the adversarial environment.

Rochester, New York Area

Working on traffic analysis using machine learning, deploying deep neural networks as website fingerprinting attacks on encrypted traffic. We developed a website fingerprinting attack against Tor network using Convolution Neural Networks (CNN). A Website fingerprinting attack is machine learning problem that enables a local eavesdropper to determine which websites a user is visiting over an encrypted connection. Our CNN attack attains over 98% accuracy on Tor traffic without defenses, better than all prior attacks, and it is also effective against lightweight website fingerprinting defenses in Tor.

Ph.D., Computer Science

Master'S Degree, Electrical And Electronics Engineering