1 code implementation • 6 Mar 2024 • Dario Pasquini, Martin Strohmeier, Carmela Troncoso
We introduce a new family of prompt injection attacks, termed Neural Exec.
no code implementations • 12 Feb 2024 • Sumeet Ramesh Motwani, Mikhail Baranchuk, Martin Strohmeier, Vijay Bolina, Philip H. S. Torr, Lewis Hammond, Christian Schroeder de Witt
In this paper, we comprehensively formalise the problem of secret collusion in systems of generative AI agents by drawing on relevant concepts from both the AI and security literature.
1 code implementation • 24 Oct 2022 • Christian Schroeder de Witt, Samuel Sokota, J. Zico Kolter, Jakob Foerster, Martin Strohmeier
Steganography is the practice of encoding secret information into innocuous content in such a manner that an adversarial third party would not realize that there is hidden meaning.
no code implementations • 23 Nov 2021 • Christian Schroeder de Witt, Yongchao Huang, Philip H. S. Torr, Martin Strohmeier
We then argue that attacker-defender fixed points are themselves general-sum games with complex phase transitions, and introduce a temporally extended multi-agent reinforcement learning framework in which the resultant dynamics can be studied.
1 code implementation • 17 Jul 2021 • Samuel Sokota, Christian Schroeder de Witt, Maximilian Igl, Luisa Zintgraf, Philip Torr, Martin Strohmeier, J. Zico Kolter, Shimon Whiteson, Jakob Foerster
We contribute a theoretically grounded approach to MCGs based on maximum entropy reinforcement learning and minimum entropy coupling that we call MEME.
1 code implementation • 8 Jul 2020 • Giulio Lovisotto, Henry Turner, Ivo Sluganovic, Martin Strohmeier, Ivan Martinovic
Research into adversarial examples (AE) has developed rapidly, yet static adversarial patches are still the main technique for conducting attacks in the real world, despite being obvious, semi-permanent and unmodifiable once deployed.
2 code implementations • 12 Feb 2020 • James Pavur, Martin Strohmeier, Vincent Lenders, Ivan Martinovic
However, status-quo services are often unencrypted by default and vulnerable to eavesdropping attacks.
Cryptography and Security Networking and Internet Architecture Performance
no code implementations • 30 Jul 2019 • Martin Strohmeier, Matthew Smith, Vincent Lenders, Ivan Martinovic
Classi-Fly obtains the correct aircraft category with an accuracy of over 88%, demonstrating that it can improve the meta data necessary for applications working with air traffic communication.