MC Mohamad Chehade
Publications
2025 ICML

LEVIS: Large Exact Verifiable Input Spaces for Neural Networks

M. Chehade, W. Li, B. W. Bell, R. Bent, S. R. Kazi, H. Zhu

Proceedings of the 41st International Conference on Machine Learning (ICML), 2025

A verification method for constructing large exact input spaces over which neural-network behavior can be certified.

paper OpenReview
LEVIS paper preview showing verifiable neural-network input regions
Preview cropped from the paper PDF.

Contribution

What this paper adds

  • Constructs verifiable input regions for neural networks.
  • Focuses on exactness and scale rather than only local certificates.
  • Connects verification with deployment-time reliability questions.