( More.)ġ990: End-to-end differentiable sequential neural attention. ( More.)ġ990: Planning & reinforcement learning with recurrent world models and artificial curiosity (1990).ġ990: Artificial Curiosity / Generative Adversarial Networks. 2021: New stuff!ġ991: First working Deep Learner based on unsupervised pre-training. The 2021 Turing Lecture, and the 2018 Turing AwardĢ022: Celebrating the 1997 LSTM paper and other works from a quarter-century agoĢ022: The most cited neural networks all build on work done in my labsġ997-: Computable universes / generalized algorithmic informationġ991: Neural nets learn to program neural nets with fast weights-Transformers with linearized self-attention.

Generalized algorithmic information & Kolmogorov complexityĢ000-: The Speed Prior-a new simplicity measure for near-optimal computable predictionsĢ021: Starting as Director of the AI Initiative at KAUST, the university with the highest impact per facultyĢ022: LeCun's paper on autonomous machine intelligence rehashes but does not cite essential work of 1990-2015Ģ022: Scientific Integrity and the History of Deep Learning: The rest is historyĢ009-: First contests won by recurrent NNs andĢ007: Deep reinforcement learning with policy gradients for LSTM.Ģ005: 1st paper with 'learn deep' in the title.Ģ003-: Gödel machines-mathematically optimal general problem solversĢ002:- Optimal Ordered Problem Solver (asymptotically optimal curriculum learner)

Mar 2012: First Deep Learner to win an image segmentation competitionĪug 2011: First superhuman visual pattern recognitionĢ011: DanNet triggers the deep convolutional neural network (CNN) revolutionĢ010: Supervised deep learning breakthrough. Sep 2012: First Deep Learner to win a medical imaging contest (cancer detection) ( More.)Ģ007-: Deep reinforcement learning with policy gradients for LSTM May 2015: Highway Networks: First working feedforward neural networks (NNs) with over 100 layersįirst deep reinforcement learning from high-dimensional sensory input. Ģ011-2015: First deep convolutional NNs to win computer vision contests (5) Additional Papers on Personal Grants.ġ990-1991: Our Miraculous Year of Deep Learningġ990-: 3 decades of artificial curiosity & creativity (and GANs)ġ991-: Transformers with linearized self-attention / Fast Weight Programmers / NNs learn to program NNsġ990-: Planning & reinforcement learning with recurrent world models and artificial curiosityġ987-: Metalearning machines that learn to learnġ991-: Very deep learning with unsupervised pre-trainingĢ010-2017: Our impact on the world's most valuable public companies: Apple, Google, Microsoft, Facebook, Amazon. Is followed by over 400 peer-reviewed publications listed in 5 Sections on

Ai technical publications pdf