A

PRIMA: general and precise neural network certification via scalable convex hull approximations

Proceedings of the ACM on Programming Languages

Authors 0

  1. Author list not loaded yet.

Cited by 2 stored of 75

2 results

References 0