MLP-C++

Overview

This repository hosts a Multi-Layer Perceptron (MLP) implemented from scratch in C++. It was developed as a class project for a Master's level "Programming for Engineers" course.

What is a Multi-Layer Perceptron (MLP)?

A Multi-Layer Perceptron (MLP) is a type of artificial neural network composed of multiple layers of interconnected nodes or "neurons". It is an extension of the simple perceptron, which consists of a single layer of neurons. MLPs can approximate non-linear functions, making them versatile for various tasks.

Input Layer: Where the model receives input from datasets.

Hidden Layers: Intermediate layers where calculations take place.

Output Layer: Where the final prediction or classification is made.

Through a process called "backpropagation" the network adjusts its weights based on the error of its predictions.

Significance in AI

MLPs are foundational to deep learning, which has powered a variety of advancements in AI. Here are some areas that have been influenced by the concepts of MLPs:

Deep Neural Networks (DNNs): MLPs can be thought of as early DNNs with multiple layers. DNNs expand on this by having many hidden layers, making them "deep".

Convolutional Neural Networks (CNNs): Used primarily in image processing, CNNs build upon the layered structure of MLPs but add specialized layers for tasks like feature extraction.

Recurrent Neural Networks (RNNs): Applied mainly for sequence data like time series or natural language, RNNs are influenced by MLPs but add loops to retain information.

I embarked on this project to deepen my understanding of Neural Networks and their underlying optimization techniques.

Features

Implmentation of backpropagation for weight adjustment.

Implementation of dropout for regularization.

Batch normalization for improved convergence.