Caner Tol



I am a Ph.D. candidate at Vernam Lab, WPI, advised by Berk Sunar.

My current research interests are

  • Microarchitectural (ÎĽarch) security
  • Side-channel and Fault Injection attacks
  • Attacks on ML models and other critical software
  • Automated detection and patching of ÎĽarch vulnerabilities using AI

Outside of work, I do Aikido, scuba diving, sailing, tango…


Nov 16, 2023 Our Mayhem attack has been accepted to ACM AsiaCCS 2024, Singapore!
Sep 25, 2023 We have issued CVE-2023-42465 for the SUDO library for our Mayhem attack. A series of commits were pushed as countermeasures.
Apr 23, 2023 I’ve been selected to receive a student travel grant for IEEE S&P 2023.
Apr 13, 2023 Our DNN backdoor injection paper is accepted to IEEE DSN 2023, Porto, Portugal.
Mar 10, 2023 Our Jolt attack has been accepted to IEEE S&P 2023, San Francisco, CA!
Dec 12, 2022 We have issued three CVEs for Amazons2n (CVE-2022-42962), WolfSSL (CVE-2022-42961), LibreSSL (CVE-2022-42963) libraries for our Jolt attack.
May 23, 2022 I’ve started my summer internship at NVIDIA as an Hardware Security Intern. :sparkles:
Dec 4, 2019 I have attended Microsoft Private AI Bootcamp along with 30 other PhD students around the world.

selected publications

  1. ZeroLeak: Using LLMs for Scalable and Cost Effective Side-Channel Patching
    M. Caner Tol, and Berk Sunar
  2. Mayhem: Targeted Corruption of Register and Stack Variables
    Andrew J. Adiletta, M. Caner Tol, Yarkın Doröz, and Berk Sunar
    In Proceedings of the 2024 ACM Asia Conference on Computer and Communications Security 2023
  3. Don’t Knock! Rowhammer at the Backdoor of DNN Models
    M. Caner Tol, Saad Islam, Andrew J. Adiletta, Berk Sunar, and Ziming Zhang
    In Proceedings of the 2023 IEEE/IFIP International Conference on Dependable Systems and Networks (DSN) 2022