# A Universal Logic Operator for Interpretable Deep Convolution Networks

## TL;DR

This paper introduces a universal logical operator for deep convolutional networks that enhances interpretability by learning logical operations automatically, moving beyond predefined AND, OR, XOR functions.

## Contribution

It proposes a novel method to learn a universal logical operator, enabling more interpretable neural networks without manual specification of logical functions.

## Key findings

- Provides a new logical interpretation for convolutional networks.
- Enables automatic learning of logical operators within the network.
- Improves interpretability of deep models through logical operations.

## Abstract

Explaining neural network computation in terms of probabilistic/fuzzy logical operations has attracted much attention due to its simplicity and high interpretability. Different choices of logical operators such as AND, OR and XOR give rise to another dimension for network optimization, and in this paper, we study the open problem of learning a universal logical operator without prescribing to any logical operations manually. Insightful observations along this exploration furnish deep convolution networks with a novel logical interpretation.

---
Source: https://tomesphere.com/paper/1901.08551