CheckINN: Wide Range Neural Network Verification in Imandra (Extended)
Desmartin, Remi, Passmore, Grant, Komendantskaya, Ekaterina, Daggitt, Matthew
–arXiv.org Artificial Intelligence
Neural networks are increasingly relied upon as components of complex safety-critical systems such as autonomous vehicles. There is high demand for tools and methods that embed neural network verification in a larger verification cycle. However, neural network verification is difficult due to a wide range of verification properties of interest, each typically only amenable to verification in specialised solvers. In this paper, we show how Imandra, a functional programming language and a theorem prover originally designed for verification, validation and simulation of financial infrastructure can offer a holistic infrastructure for neural network verification. We develop a novel library CheckINN that formalises neural networks in Imandra, and covers different important facets of neural network verification.
arXiv.org Artificial Intelligence
Jul-28-2022
- Country:
- North America
- United States
- Texas > Travis County
- Austin (0.04)
- New York > New York County
- New York City (0.04)
- Illinois > Cook County
- Chicago (0.04)
- California > Los Angeles County
- Long Beach (0.04)
- Texas > Travis County
- Canada > British Columbia
- United States
- Europe
- Denmark (0.04)
- United Kingdom
- Scotland > City of Glasgow
- Glasgow (0.04)
- England > Cambridgeshire
- Cambridge (0.04)
- Scotland > City of Glasgow
- Italy > Lazio
- Rome (0.04)
- Germany > Baden-Württemberg
- Karlsruhe Region > Heidelberg (0.04)
- France > Île-de-France
- Asia
- Japan > Kyūshū & Okinawa
- Kyūshū > Fukuoka Prefecture > Fukuoka (0.04)
- Georgia > Tbilisi
- Tbilisi (0.05)
- Japan > Kyūshū & Okinawa
- North America
- Genre:
- Research Report (1.00)
- Technology: