AlexNet: The Deep Learning Breakthrough That Changed Computer Vision

Author(s): Kshitij Darwhekar

Originally published on Towards AI.

This article delves into AlexNet’s journey, from its groundbreaking architecture and innovations to its lasting impact on the field of deep learning. Explore the key features, techniques to reduce overfitting, and its legacy in shaping modern neural networks.

This member-only story is on us. Upgrade to access all of Medium.

Don’t have a paid Medium membership (yet)? You can read the entire article for free by clicking here with my friend’s link.

AlexNet named after first author Alex, was introduced in 2012. The paper titled “Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton, ImageNet Classification with Deep Convolutional Neural Networks” was published at NIPS in Sept 2012. Most of the object recognition models at that time used essential machine learning methods.

The authors noticed that the performance of the models, can be improved by collecting larger datasets, using better techniques to reduce overfitting. They realized that to train thousands of objects from millions of images, they needed large learning capacity. But due to immense complexity of object recognition this task won’t be achieved even with a large dataset like ImageNet. To solve this, they needed a large model with lots and lots of prior knowledge.

CNNs were the obvious choice due to nature of the complexities explained earlier also the capacity of CNNs could be easily controlled by varying depth and breadth. Apart from this CNNs used to make strong and mostly correct assumption about nature of images. Despite of all the attractive… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI