Category: Research
-
DARPA Grand Challenge 2005: Stanley crosses the desert (2005)
On 8 October 2005 Stanford's autonomous vehicle Stanley won the DARPA Grand Challenge, completing a 212 km Mojave Desert course in 6 hours 53 minutes after no vehicle finished the 2004 race.
-
LeCun's zip code nets: when convolution met backpropagation (1989)
In 1989 Yann LeCun and colleagues at AT&T Bell Laboratories published a convolutional neural network trained with backpropagation on handwritten US Postal Service zip codes, an early proof that local filters could scale to real mail sorting.
-
Neural language models: Bengio's fight against the curse of dimensionality (2003)
In 2003 Yoshua Bengio and colleagues published A Neural Probabilistic Language Model in the Journal of Machine Learning Research, learning distributed word representations to escape the limits of huge n-gram tables.
-
Random forests: Leo Breiman's ensemble of decision trees (2001)
In October 2001 Leo Breiman published Random Forests in Machine Learning, combining many decorrelated decision trees trained on bootstrap samples to create a powerful general-purpose classifier.
-
Support vector machines: maximum-margin classifiers with kernels (1995)
In 1995 Corinna Cortes and Vladimir Vapnik published support-vector networks in Machine Learning, combining maximum-margin linear separation with the kernel trick for nonlinear boundaries.
-
TD-Gammon: learning backgammon by playing itself (1992)
Gerald Tesauro at IBM built TD-Gammon, a neural network that learned strong backgammon through temporal-difference reinforcement learning and self-play, reaching near expert-level play.
-
Boltzmann machines: when neural networks borrowed statistical physics (1985)
In 1985 David Ackley, Geoffrey Hinton and Terrence Sejnowski published a learning algorithm for Boltzmann machines: stochastic binary units, an energy function from physics, and slow sampling that later methods would try to avoid.
-
Deep belief networks: when deep nets became trainable again (2006)
In July 2006 Geoffrey Hinton and colleagues showed that stacking restricted Boltzmann machines layer by layer, then fine-tuning with backpropagation, could train deep networks. The result helped restart the deep-learning revival.
-
Backpropagation: how neural networks learned from their mistakes (1986)
In 1986 David Rumelhart, Geoffrey Hinton and Ronald Williams showed that multilayer neural networks trained with backpropagation learn their own internal representations. The method is still the standard way to train most neural networks.
-
GANs: when neural networks learned to invent images (2014)
In June 2014 Ian Goodfellow and colleagues introduced generative adversarial networks: two neural networks trained against each other. The idea made realistic image synthesis possible and raised the question of fakes early on.
-
AI reaches gold at the International Mathematical Olympiad (2025)
In July 2025, language models from Google DeepMind and OpenAI solve five of six IMO problems for 35 of 42 points, writing proofs in natural language within the exam time limit.
-
AlphaFold wins the 2024 Nobel Prize in Chemistry
One day after AI took the physics prize, the 2024 Chemistry Nobel honored AlphaFold, the DeepMind model that predicts protein structures, and computational protein design.
-
The 2024 Nobel Prize in Physics goes to neural networks
John Hopfield and Geoffrey Hinton won the 2024 Nobel Prize in Physics for foundational work on artificial neural networks, the machinery behind today's AI.
-
"Attention is all you need": the Transformer arrives (2017)
A 2017 paper replaced the sequential machinery of earlier networks with pure attention. The Transformer became the architecture behind almost every large language model since.
-
AlexNet: the night deep learning came back (2012)
A deep neural network trained on two gaming <abbr title="Graphics processing unit">GPU</abbr>s crushed the ImageNet benchmark and ended a long winter for neural networks almost overnight.
-
GPT-3: when scale became the strategy (2020)
OpenAI's 175-billion-parameter model showed that a big enough language model could learn new tasks from a few examples in the prompt, no retraining required.
-
ImageNet: the dataset that would teach machines to see (2006)
Fei-Fei Li bet that data, not just cleverer algorithms, would unlock computer vision. ImageNet became the benchmark that lit the fuse for the deep-learning era.
-
OpenAI o1: models that think before they answer (2024)
OpenAI previewed a model line trained to reason through problems step by step before responding, trading speed for accuracy and reviving the idea that thinking longer can pay off.
-
Rosenblatt's perceptron: the machine that learned (1958)
Frank Rosenblatt's perceptron was the first machine to learn from experience, celebrated by the press, attacked by critics, and vindicated decades later.
-
Computing Machinery and Intelligence: Turing's imitation game (1950)
Alan Turing's 1950 paper replaced 'Can machines think?' with a practical test, and shaped how we still talk about machine intelligence.