Generative Adversarial Networks
Goodfellow et al., 2014 — proposed training two networks against each other, a generator and a discriminator, as a way to learn to generate realistic data. Dominated image generation for most of the following decade.
"Generative Adversarial Networks" (Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, and colleagues, 2014) introduced the GAN, the first widely adopted way to train a network to generate realistic data without writing down an explicit probability model for it. The paper's own samples (MNIST digits, faces, small CIFAR-10 images) were modest by later standards; the photorealism came from the follow-up work that built on it.
What problem it solved
Before this paper, generative models — ones that produce new data rather than classify existing data — mostly relied on explicitly modeling a probability distribution over the data and sampling from it, an approach that scaled poorly to complex, high-dimensional data like natural images. There was no training setup that let a network implicitly learn to produce realistic samples without first solving the much harder problem of writing down an explicit density function for what "realistic" meant.
The key idea
Turn generation into a two-player game instead of a direct optimization problem. Train two networks simultaneously, with opposing goals:
- The generator takes random noise as input and tries to produce outputs realistic enough to fool the discriminator.
- The discriminator is an ordinary classifier, trained to tell the generator's fake outputs apart from real training examples.
Each network's training signal comes from the other: the generator improves by getting better at fooling an increasingly capable discriminator, and the discriminator improves by getting better at catching an increasingly capable generator. Neither network needs an explicit model of what "realistic" means — that entire problem is outsourced to the adversarial game itself, which the paper formalizes as a minimax objective with a Nash equilibrium as its target rather than a simple minimum.
Why it mattered
GANs became the dominant approach to high-fidelity image generation for roughly the next decade, producing results — photorealistic faces, style transfer, image-to-image translation — that earlier generative approaches couldn't match. The adversarial training setup itself proved fragile (mode collapse, unstable convergence, no single loss to monitor), which is a large part of why diffusion modelsDiffusion ModelA diffusion model learns to reverse a fixed process of gradually adding noise to data, generating new samples by denoising pure noise step by step. eventually displaced GANs for state-of-the-art image generation — but the core idea, that a learned discriminator can substitute for a hand-designed loss function, outlived the specific architecture and reappears across generative modeling more broadly.
Authors: Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, Yoshua Bengio (Université de Montréal)
Read the paper — arXiv:1406.2661See also: Ian GoodfellowIan GoodfellowInvented generative adversarial networks (GANs) in 2014, the architecture that first made realistic image generation possible by pitting two networks against each other. · Yoshua BengioYoshua BengioHelped lay the statistical foundations of modern language modeling and co-founded the Montreal deep learning school — now a leading voice on AI safety.
Learn more: GANGAN (Generative Adversarial Network)A GAN trains a generator and a discriminator against each other, the generator learning to fool the discriminator into mistaking its output for real data. · Generative Models
Learning Transferable Visual Models From Natural Language Supervision
Radford et al., 2021 (the "CLIP" paper) — trained an image encoder and a text encoder together on 400 million image-caption pairs, producing a shared embedding space that later became the standard bridge into multimodal models.
Auto-Encoding Variational Bayes
Kingma & Welling, 2013 — introduced the VAE, showing how to train an encoder-decoder pair so that its compressed latent space is smooth and sampleable enough to generate new, plausible data.