Demis Hassabis

∙

0 followers

Demis Hassabis is a British artificial intelligence investigator, neuroscientist, video game designer, entrepreneur and internationally recognized gambler.

Hassabis was born to a father from Greece and Cyprus and a mother of Chinese Singapore. Hassabis, the child prodigy of Chess, reached the master standard at the age of 13 with an Elo rating of 2300. He represented Cambridge University in the 1995, 1996 and 1997 varsity chess matches in Oxford-Cambridge and won a half-blue match.

Hassabis was educated at Christ’s College, Finchley, East Finchley, North London, a state-funded full school. At 15 and 16 years old, he completed his GCE Advanced Level and Scholarship Level exams.

Hassabis started his career with the computer games at Bullfrog Productions, first level with the game designer Peter Molyneux and 17 co-designers with the game designer Theme Park in 1994. The Theme Park, a simulation video game, has sold several million copies and won the Golden Joystick Award.

He then left Bullfrog for Queens’ College, Cambridge, where he graduated with the Double First from the University of Cambridge in 1997 with the Computer Science Tripos.

Hassabis studied in Lionhead Studios after graduating from Cambridge. Recent company founded by renowned game designer Peter Molyneux, with whom Hassabis had been involved at Bullfrog Productions. Hassabis worked as a lead AI programmer at Lionhead for the 2001 Black & White “God” game.

In 1998 Hassabis left Lionhead to found Elixir Studios, a London-based independent developer of games, which signs Eidos Interactive, Vivendi Universal and Microsoft publishing deals. In addition to managing the company, Hassabis served as executive designer of BAFTA-nominated Games Republic: the Revolution and Evil Genius. Republic: the Revolution, Elixir’s first game, was delayed due to the huge range of its products. The final game, which was reduced by its original vision, received a metacritical score of 62/100 with tender reviews. Evil Genius, a Bond villain tongue-in-cheek simulator, was much better with a 75/100 score. In April 2005, intellectual property and technology rights were sold and the studio closed to various publishers.

In 2009, after the Elixir Studios, Hassabis returned to academia and was supervised by Eleanor Maguire to obtain his doctorate in cognitive neuroscience at University College London. He sought to find inspiration in the human head for new AI algorithms. As a visiting scientist at the Massachusetts Institute of Technology and Harvard University he continued his research on neuroscience and artificial intelligence before earning an Henry Wellcome postdoctoral fellowship in 2009 at the Gatsby Charitable Foundation Computational neuroscience Unit, UCL One of his highly cited papers in the Proceedings of the National Academy of Science of the USA systematically showed for the first time that patients with amnesias-causing damage to their hippocampus were also unable to imagine themselves in new experiences. The finding established a connection between the constructive imagination process and the reconstructive process of episodic recall. Hassabis has created a new theoretical report of the Episodic Memory System which identifies the scene construction, generation and online maintenance of a complex and coherent scene as a major process that underlies memory remembrance and imagination. On this basis, Hassabis studied functional magnetic resonance imaging. This piece was widely covered in the mainstream media and listed in any field by the journal Science as the top 10 scientific breakthroughs of the year.

In 2010, Hassabis co-founded DeepMind with Shane Legg and Mustafa Suleyman, a London based AI machine learning firm. Hassabis and Suleyman had been friends since infancy, and he met Legg when they both were postdocs at the Gatsby Computational Neuroscience Unit of University College, London. The mission of Hassabis is to “solve intelligence” and use intelligence “to resolve all else.” Hassabis also recruited his University friende and Elixir partner David Silver. More specifically, DeepMind is seeking to unlock increasingly powerful general purpose learning algorithms aimed at the development of artificial general intelligence through new developments in computer hardware. The company has focussed on training learning algorithms for master games and, in December 2013, it was famous to announce a groundbreaking breakthrough through training of an algorithm called a Deep Q-Network for superhuman Atari games only with the use of raw pixels on screen. In 2014, Google bought DeepMind for $400 million, even though it remained an independent company based in London. The company has recorded a number of important achievements since it acquired Google, perhaps the most remarkable being the creation of AlphaGo, a program that has defeated world champion Lee Sedol in the complex game Go. For the high number of possible board positions and resistance to already existing programming techniques, Go was considered a holy grail of AI. In October 2015, however, AlphaGo beat European champion Fan Hui 5-0 before winning 4-1 against former Lee Sedol in March 2016. Further achievements of DeepMind include creating a Neural Turing Machine, promoting AI safety research, and creating a partnership with the United Kingdom National Health Service and Moorfields Eye Hospital to improve medical services and to detect the development of degenerative conditions in the eye. The 13th Critical Evaluation of Protein Structure Prediction Techniques was won by DeepMind’s AlphaFold in December 2018 by successfully predicting the most accurate structures for 25 of 43 proteins. “It is a lighthouse project, our first major investment in people and resources into a fundamental and very important scientific problem in the real world,” Hassabis told the Guardian. This demonstration of the power of AI, specifically the Application by DeepMind, provides a good insight into causes and cures of diseases and serves as a component of potential Noble Price breakthroughs over the next decade. DeepMind also made a number of award-winning papers responsible for technical improvements in machine learning. The company has, in particular, made substantial progress in profound learning and strengthened learning and pioneered the field of profound strengthening learning that combines these two methods. Hassabis has predicted that Artificial Intelligence will “be one of the most advantageous technologies in the human race.”

Top 10 scientific breakthroughs of 2007 Science Magazine

Royal Arts Society Fellow 2009 2009

Henry Wellcome Fellowship in Postdoctoral Research

Queens’ College, Cambridge Fellow Benefactor

Mullard Royal Society Award

London Evening Standard’s third most influential Londoner in 2014

Listed on ‘Smart 50’ Wired

Financial Times Top 50 European Entrepreneurs

Digital entrepreneur of the year Financial Times

University College London Honorary Fellow

London Evening List of influential Londoners, No. 6

Silver Medal Royal Academy of Engineering

WIRED Innovation Leadership

“Ten people of nature who mattered this year”

2016 Top 10 Scientific Breakthroughs Science Magazine

Time 100: The 100 Most Important People

The Asian Awards: Excellence in Science and Technology

Royal Academy of Engineering Fellow

Academy of Realization: Golden Plate Award

Appointed Commander of the Order of the British Empire for ‘Science and Technology Services’ in 2018.

Imperial College London awarded an honorary degree in May 2018.

Elected Royal Society Fellow in May 2018

Computer Laboratory Company of the Year Cambridge

Two articles in front of Nature

9-Dan Go Honorary AlphaGo Rank of the Baduk Association of Korea

Grand Prix Cannes Lion for AlphaGo

WIRED Award for Innovation in AI

City AM Year’s innovative company

Hassabis is a skilled player in many games, including:

Chess: at the age of 13 with ELO rating of 2300 achieved the Master standard.

Diplomacy: World Team Champion 2004, 4th World Championship 2006, 3rd European Championship 2004.

Poker: cashed six times at the World Poker Series including the Main Event.

World Pentamind Champion and World Champion Decamentathlon Multi-game events at the London Mind Sport Olympiad.