\title{Trying to Prove Neural Networks' Equivalence \newline (and other cool properties)}
\author{Charis Eleftheriadis
\and
Panagiotis Katsaros\\
\and
Stavros Tripakis
}
\date{February 11, 2021}
\begin{document}
\maketitle
This is a living document comprised of thoughts, findings, paradigms regarding \textbf{Neural Networks' Equivalence}.
\vspace{5mm}
\newline At the moment, we try to use SAT, SMT, MIP solvers to prove various properties of Neural Networks.
Our main research goal is to provide a formal way to train provable robust Neural Networks of different complexity and size (number of neurons and network's parameters).