Karan Diwan

Artificial Neuron

A gate-level McCulloch–Pitts neuron: the smallest bridge from logic gates to a model of a neuron.

This project implements a basic McCulloch–Pitts artificial neuron directly in Verilog: two binary inputs, a programmable threshold, and inhibitory weights. Varying the threshold alone is enough to reproduce AND and OR logic; adding an inhibitory weight reproduces NOT. It marks the point where the rest of this work turns from pure arithmetic toward models of learning and decision, continued in the Gate-Level Cybernetic Classifier.

McCulloch-Pitts neuron block diagram
MCP neuron reproducing AND and OR logic via threshold adjustment.

Versions built

RTL waveform of the neuron
RTL simulation waveform, MCP-Neuron v0.
Toolchain: Verilog, Icarus Verilog, Yosys, GTKWave, Python (reference model)
View repository on GitHub →
← Assembler