September 2012

AlexNet wins ImageNet

Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton submitted AlexNet to ILSVRC 2012 on September 30, 2012 — winning with 15.3% top-5 error and triggering the deep learningNeural networks with many layers that learn hierarchical representations from data. revolution.

What it was for

AlexNetAlexNet's GPUGraphics processing unit — a chip optimized for parallel math; now used for graphics, ML, and scientific computing.-trained convolutional network crushed traditional computer visionSoftware that interprets images and video — object detection, OCR, and scene understanding. pipelines — convincing Google, FacebookThe social network founded by Mark Zuckerberg — grew into Meta and reshaped online advertising., and industry to bet on neural networkA model of connected layers that learns from examples — the core of modern deep learning.s. Hinton had pursued backprop for decades before this breakthrough.

People

  • Geoffrey Hintonresearcher

Why it's here

AlexNet was the inflection point for modern deep learningNeural networks with many layers that learn hierarchical representations from data. adoption.

Why it mattered

It proved neural networkA model of connected layers that learns from examples — the core of modern deep learning.s could beat hand-engineered vision features at scale.

What it solved

Computer visionSoftware that interprets images and video — object detection, OCR, and scene understanding. relied on brittle handcrafted features that did not generalize.

Media

  • AlexNet
    ImageAlexNet

    Zhang, Aston and Lipton, Zachary C. and Li, Mu and Smola, Alexander J., CC BY-SA 4.0, via Wikimedia Commons

Related