Neural Network Research with Wolfram Language: Exploring Non-Differentiable Boolean Functions

Explore cutting-edge neural network research using Wolfram Language in this 29-minute talk. Discover how to leverage advanced language features for rapid prototyping of neural network applications and innovative research ideas. Learn about the development of a novel neural network capable of learning compact non-differentiable Boolean functions. Delve into topics such as relaxations, margin packing, differentiable majority, and hard NetCode. Gain insights into the problem-solving process, implementation techniques, and the impressive results achieved through this innovative approach to Boolean function learning.

Introduction
The Problem
The Basics
Relaxations
Margin Packing
Differentiable Majority
Hard Net
Code
Boolean function
Results