My Journey Into Learning About ML and Implementing a MLP From Scratch
Note: This is my first ever blog post, so my apologies in advance if it's not up to par. Also, since this is my first post, I decided to not use any AI for polishing, etc., so that the reader can get a better feel for who I am as a writer. Introduction The purpose of this blog post is to document my journey, as a current non-professional or hobbyist, into the world of Artificial Intelligence,…
This blog post documents the author's journey into the world of Artificial Intelligence, specifically Machine Learning (ML). The author, a non-professional hobbyist, enrolled in two Nanodegree programs through Udacity to learn the concepts and math behind ML. After completing these programs, the author decided to delve deeper by learning a new programming language and implementing the ML algorithms from scratch on both CPU and GPU.
The author chose Julia as the programming language for its recent emergence as a high-level language. They focused on implementing a Multi-Layer Perceptron (MLP) instead of an advanced Neural Network due to the complexity of derivatives in convolutional operations.
The author emphasizes the importance of understanding the math behind ML algorithms and spent considerable time working out the derivatives necessary for the back-propagation (BP) step for both batch sizes of 1 and greater than 1. While the feed-forward step was relatively simple, the BP step required significant effort.
The implementation phase involved writing functions for the math, organizing the code for efficiency and readability, and grappling with the patience required to find the right hyperparameters for good results. The author spent many hours and training cycles to achieve a 99% accuracy on the training dataset and under a 5-minute training time, surpassing their initial goals.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.