Search papers, labs, and topics across Lattice.
University of Wisconsin-Madison
4
0
7
Agentic AI could revolutionize cybersecurity by transforming labor-intensive tasks into efficient, automated defenses.
Achieving sound probabilistic verification for AI agents could redefine how we secure complex systems against policy violations in uncertain environments.
Provably undetectable backdoors can be injected into pre-trained image classifiers, even with white-box access, by exploiting sparse perturbations and Gaussian dithering.
Forget prompt engineering: PCAS deterministically enforces complex authorization policies in multi-agent systems, boosting compliance from 48% to 93% without restructuring existing agents.
