The source code for AlexNet - the neural network developed at the University of Toronto that kickstarted today's artificial intelligence boom and led to a Nobel Prize - will be preserved by the Computer History Museum in partnership with Google.
The museum, located in Mountain View, Calif, boasts a diverse archive of software and related material and aims to "decode technology - the computing past, digital present, and future impact on humanity."
It has already released other historic source codes, including APPLE II DOS, IBM APL, Apple MacPaint and QuickDraw, Apple Lisa and Adobe Photoshop.
"This code underlies the landmark paper ImageNet Classification with Deep Convolutional Neural Networks by Alex Krizhevsky, Ilya Sutskever and Geoffrey Hinton , which revolutionized the field of computer vision and is one of the most cited papers of all time," says Jeff Dean, chief scientist, Google DeepMind and Google Research, of AlexNet.
"Google is delighted to contribute the source code for the groundbreaking AlexNet work to the Computer History Museum."
AlexNet has its roots in the decades of research conducted by Hinton, a U of T University Professor Emeritus of computer science who recently shared the 2024 Nobel Prize in Physics with Princeton's John Hopfield for foundational work in AI.
By the early 2000s, Hinton's graduate students at U of T were beginning to use graphics processing units (GPUs) to train neural networks for image recognition tasks and their success suggested that deep learning could be a path to creating general-purpose AI systems.
In particular, Sutskever - who went on to become a key figure at OpenAI, which launched ChatGPT, and will receive an honorary degree from U of T this year - believed that the performance of neural networks would scale with the amount of data available.
The arrival of ImageNet in 2009 provided him with the chance to test his theory. The dataset of images developed by Stanford University Professor Fei-Fei Li was larger than any previous image dataset by several orders of magnitude.
In 2011, Sutskever convinced Krizhevsky, a fellow graduate student, to train a convolutional neural network on ImageNet. With Hinton serving as principal investigator, Krizhevsky programmed the network on a computer with two NVIDIA cards. Over the course of the next year, he tweaked the network's parameters and retrained it until it achieved performance superior to its competitors.
The network was ultimately named AlexNet in his honour.
Before AlexNet, very few machine learning researchers used neural networks. After it, almost all of them would. Google eventually acquired the company started by Hinton, Krizhevsky and Sutskever, and a Google team led by David Bieber worked with CHM for five years to secure the code's public release.
In describing the AlexNet project, Hinton says, "Ilya thought we should do it, Alex made it work and I got the Nobel Prize."
With files from the Computer History Museum
