2014

Ian Goodfellow

b. 1985

Generative Adversarial Networks

Introduced generative adversarial networks, in which two competing neural networks learn to create increasingly convincing synthetic examples.

In 2014, Goodfellow and colleagues proposed GANs: a generator produces samples while a discriminator learns to distinguish them from real data. Their competition gives the generator a powerful learning signal.

GANs became influential in image synthesis, editing and research on generative models. Goodfellow also co-authored the Deep Learning textbook with Yoshua Bengio and Aaron Courville, helping establish the field's shared foundations.

Dated contributions

  1. 2014

    Generative Adversarial Networks

    Ian Goodfellow

    Introduced generative adversarial networks, in which two competing neural networks learn to create increasingly convincing synthetic examples.

    Paper · 2014 — Generative Adversarial Nets ↗

Reference

Original work & historical records

Wikipedia