3 papers indexed
Andrew C. Cullen, Neil G. Marchant, Jiani Xie, Paul Montague, Benjamin I. P. Rubinstein
Automatic Speech Recognition systems are notoriously both sensitive to adversarial and benign perturbations. While this has been repeatedly demonstrated using reference datasets, detecting such behaviors in deployed systems is incredibly challenging, due to the absence of oracle…
Andrew C. Cullen, Neil G. Marchant, Jiani Xie, Paul Montague, Sean Lamont, Maxwell Standen, et al.
While voice control is rapidly becoming a ubiquitous vector of human-AI communication, the risks facing these systems remain poorly understood. This is, in part, a product of the difficulties in scaling strictly digital adversarial workflows to the physical world. These scale bar…
Andrew C. Cullen, Paul Montague, Benjamin I. P. Rubinstein
Randomized Smoothing (RS) provides rigorous robustness guarantees for neural networks without architectural constraints, yet its adoption is limited by extreme computational costs. Standard RS requires tens of thousands of model evaluations per input and forces practitioners to c…