Geoffrey Hinton, Yann LeCun and Yoshua Bengio are considered the ‘godfathers’ of an essential technique in artificial intelligence, called ‘deep learning’, which is based on the use of neural networks for voice recognition, computer vision and natural language processing, and has enabled advances in fields as diverse as object perception and machine translation. These neural networks aim to mimic the functioning of the human brain, using algorithms that convert the biological process of learning into mathematical sequences. The idea is for the machine to learn from its own experience. In 1986, Hinton invented backpropagation algorithms, fundamental for training neural networks. In 2012, these algorithms allowed him to create a convolutional neural network called AlexNet, made up of 650 000 neurons and trained with 1.2 million images, which registered an error rate in object recognition of only 26%, halving that of previous systems. He has made other contributions to artificial neural networks and their training, such as the co-invention of the Boltzmann machine, the Helmholtz machine and the so-called Product of Experts (PoE). In 2021, he published a document on the arXiv preprint platform in which he presented GLOM, an innovative project as yet theoretical which involves a new vector model for processing and representing visual information in a neural network, which is still in the development phase. Yann LeCun in turn made contributions to the development of the backpropagation algorithms that Hinton had invented and in 1989 created LeNet-5, a recognition system for characters written on bank checks, which represented a major advance for optical character recognition technology. He later contributed to the development of DjVu image compression technology, used by hundreds of websites and millions of users to access scanned documents on the Internet. He has also worked on deep learning methods for document recognition, human-computer interaction and speech recognition.
For his part, Bengio has made key contributions to probabilistic sequence models, used for speech and handwriting recognition and in unsupervised learning. He is currently studying more efficient algorithms in data representations, extracting pattern recognition and also enabling the understanding of more complex relationships and high-level concepts. Demis Hassabis is CEO and co-founder of DeepMind, one of the largest artificial intelligence research companies in the world, founded in 2011 and acquired in 2014 by Google (2008 Prince of Asturias Award for Communication and Humanities). At DeepMind, Hassabis has created a neural network model that combines the capabilities of an artificial neural network with the algorithmic power of a programmable computer. Hassabis’ company has combined the advances made in machine learning with deep learning processes and what is known as reinforcement learning to devise a new field of deep reinforcement learning, an artificial intelligence system that opens the door to multiple applications in the fields of many scientific disciplines. In 2021, the DeepMind team managed to predict the structure of more than 350 000 human proteins (44% of all known proteins) with a very high degree of accuracy. The data was made available to all laboratories in the world via the AlphaFold Protein Structure Database and the achievement was featured by Science magazine (2007 Prince of Asturias Award for Communication and Humanities) as Scientific Discovery of the Year. Edith Heard, director of the European Molecular Biology Laboratory, declared that the achievement “is truly a revolution for the life sciences, just as genomics was several decades ago.” Hinton, LeCun and Bengio were distinguished in 2018 with the Turing Award granted by the Association for Computing Machinery (ACM).
Geoffrey Hinton
Geoffrey Hinton (London, UK, 6th December 1947) graduated in Experimental Psychology from the University of Cambridge in 1970 and received his PhD in Artificial Intelligence from the University of Edinburgh in 1975. He has worked at the University of Sussex (UK), the University of California, San Diego and Carnegie-Mellon University (USA), as well as at the University of Toronto (Canada), among others. Between 1998 and 2001, he set up the Gatsby Computational Neuroscience Unit at University College London. He subsequently returned to the University of Toronto, where he is currently Professor Emeritus in the Department of Computer Science. Since 2013, he has worked with Google as vice president for the development of deep learning applications and he is currently Chief Scientific Advisor at the Vector Institute of Canada. He is the author or co-author of more than four hundred publications, has been cited 596 426 times and has an h-index of 170, according to Google Scholar. Hinton is a Fellow of the Royal Society of the United Kingdom (2011 Prince of Asturias Award for Communication and Humanities), the Royal Society of Canada and the Association for the Advancement of Artificial Intelligence (of international scope). He is an honorary member of the American Academy of Arts and Sciences and the US National Academy of Engineering. Companion of the Order of Canada and holder of honorary degrees from the Universities of Edinburgh, Sussex and Sherbrooke, he has received, among other distinctions, the David E. Rumelhart Award (USA, 2001), the Award for Research Excellence from the International Joint Conferences on Artificial Intelligence Organization (2005), the Gerhard Herzberg Canada Gold Medal for Science and Engineering from the Natural Sciences and Engineering Research Council of Canada (2010), the NEC C&C Prize (Japan, 2016), the James Clerk Maxwell Medal awarded by the Institute of Electrical and Electronics Engineers (IEEE) and the Royal Society of Edinburgh (2016), the BBVA Frontiers of Knowledge Award (2017) and the aforementioned Turing Award.
Yann LeCun
Yann LeCun (Soisy-sous-Montmorency, France, 8th July 1960) graduated in Electrical Engineering from the ESIEE Paris in 1983 and received his PhD in Computer Science from the Pierre and Marie Curie University (Paris) in 1987. After postdoctoral research in Geoffrey Hinton’s group at the University of Toronto, he joined AT&T Bell Laboratories in 1988, where he became head of the Image Processing Research Department in 1996. After a brief tenure as a Fellow of the NEC Research Institute in Princeton, New Jersey, he joined New York University in 2003. He is director of AI Research at Facebook and maintains his academic activity at New York University, linked to the Center for Data Science (which he founded and headed from 2012 to 2014) and the Courant Institute of Mathematical Science. He is a member of the Advisory Board of the Institute of Pure and Applied Mathematics and of the Institute for Computational and Experimental Research in Mathematics (ICERM). LeCun is the author or co-author of more than three hundred publications, has been cited 257 140 times and has an h-index of 136, according to Google Scholar. He is a member of the US National Academy of Engineering and is the recipient of numerous awards, including the IEEE Neural Network Pioneer Award (2014), the IEEE PAMI Distinguished Researcher Award (2015) and the University of Pennsylvania Harold Pender Award (2018), in addition to the aforementioned Turing Award. He holds honorary degrees from the National Polytechnic Institute of Mexico (IPN) and the Federal Polytechnic School of Lausanne (EPFL). He was included in Wired magazine’s list of 100 Global Influencers in 2016.
Yoshua Bengio
Yoshua Bengio (Paris, France, 5th March 1964) received his Bachelor’s degree in Engineering in 1986 from McGill University (Canada), where he also earned his Master’s degree and PhD in Computer Science. He completed a postdoctoral fellowship at the Massachusetts Institute of Technology (MIT) and with LeCun’s group at AT&T Bell Laboratories. He has been a Professor in the Department of Computer Science and Operations Research at the University of Montreal since 1993. In addition to holding the Canada Research Chair in Statistical Learning Algorithms, he is the founder and scientific director of Mila, the Quebec Artificial Intelligence Institute, co-founder of the Element AI startup and advisor to several tech companies. He is also Scientific Director of the Institut de Valorisation des Données (IVADO) and co-chair, since 2019, of the Canadian Advisory Council on Artificial Intelligence. Bengio has published more than seven hundred articles, which have been cited 553 274 times and has an h-index of 207, according to Google Scholar. Officer of the Order of Canada and Chevalier of the French Legion of Honour, he is a Fellow of the Royal Society of Canada (2017). He has received distinctions such as the Prix Marie-Victorin from the Government of Quebec (2017), the Medal of the 50th Anniversary of Quebec’s Ministry of International Relations and Francophonie (2018), the Lifetime Award from the Canadian AI Association (2018), the Killam Prize and the IEEE CIS Neural Network Pioneer Award (2019). He is the author of three renowned books on deep learning and was one of the promoters of the Montreal Declaration for a Responsible Advancement of Artificial Intelligence.
Demis Hassabis
Demis Hassabis was born in London (UK) on 27th July 1976. A child prodigy in chess (he was already a renowned player at the age of thirteen) and passionate about programming, at seventeen he joined the Bullfrog Productions company as a video game designer, where he created hit games like Theme Park. He studied at the University of Cambridge, where he graduated in 1997 in Computer Science, and in 1998 founded the video game company Elixir Studios, with which he created projects for multinationals such as Microsoft and Vivendi. In 2009, he received his PhD in Cognitive Neuroscience from University College London and rounded out his studies at Harvard (USA) and MIT. In 2011, with the support of investors such as Elon Musk, he founded the AI company DeepMind Technologies, with which he began to create learning algorithms to master video games, such as AlphaGo (which in a short time managed to defeat Lee Sedol, world champion of Go, a game of Chinese origin), and to advance in the development of an artificial intelligence learning system, AlphaZero, considered by some experts to be revolutionary, as it combines human neuronal functioning and the connections between memory and imagination with machine learning mechanisms. With another of these variants, AlphaFold, he managed, as mentioned above, to transform the study of the 3D structure of proteins. Following Google’s acquisition of DeepMind in 2014, Hassabis continued to serve as its CEO. According to Google Scholar, he has been cited 84 250 times and has an h-index of 74. Commander of the Order of the British Empire, Hassabis is a Fellow of the Royal Society, the Royal Academy of Engineering and the Royal Society of Arts, all in the UK, and has received the Mullard Award from the Royal Society (2014), the Golden Plate Award from the American Academy of Achievement (USA, 2017) and the Dan David Prize (Israel, 2020).
Photographies:
Geoffrey Hinton: University of Toronto / Johnny Guatto
Yann LeCun: University of California
Yoshua Bengio: TEDxMontréal under license CC BY-NC-SA 2.0
Demis Hassabis: Royal Society under license CC BY-SA 4.0
Minutes of the jury
