Timezone: »

Synthesizing Robust Adversarial Examples
Anish Athalye · Logan Engstrom · Andrew Ilyas · Kevin Kwok

Wed Jul 11 08:40 AM -- 08:50 AM (PDT) @ A7

Standard methods for generating adversarial examples for neural networks do not consistently fool neural network classifiers in the physical world due to a combination of viewpoint shifts, camera noise, and other natural transformations, limiting their relevance to real-world systems. We demonstrate the existence of robust 3D adversarial objects, and we present the first algorithm for synthesizing examples that are adversarial over a chosen distribution of transformations. We synthesize two-dimensional adversarial images that are robust to noise, distortion, and affine transformation. We apply our algorithm to complex three-dimensional objects, using 3D-printing to manufacture the first physical adversarial objects. Our results demonstrate the existence of 3D adversarial objects in the physical world.

Author Information

Anish Athalye (MIT CSAIL)
Logan Engstrom (MIT)
Andrew Ilyas (Massachusetts Institute of Technology)
Kevin Kwok (LabSix)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors