USIA Partner Advisor Dr. Demis Hassabis

Demis Hassabis, Ph.D., CBEFRSFREngFRSA is an artificial intelligence researcher, neuroscientist, AI programmer, video game designer, entrepreneur, and world-class game player. A child prodigy in chess, he reached master standard at the age of 13 with an ELO rating of 2300, at the time one of the highest ratings for a boy of that age in the world.

He is the founder and CEO of DeepMind, a neuroscience-inspired AI company that develops general-purpose learning algorithms and uses them to help tackle some of the world’s most pressing challenges.

Since its founding in London in 2010, DeepMind has published over 200 peer-reviewed papers, five of them in the scientific journal Nature, an unprecedented achievement for a computer science lab. DeepMind’s groundbreaking work includes the development of deep reinforcement learning, combining the domains of deep learning and reinforcement learning. In 2014, DeepMind was acquired by Google and is now part of Alphabet.

Demis was born on the 27th of July 1976 to a Greek Cypriot father and a Chinese Singaporean mother and grew up in North London. A child prodigy in chess, Demis reached master standard at the age of 13 with an Elo rating of 2300 and captained many of the England junior chess teams. He represented the University of Cambridge in the Oxford-Cambridge varsity chess matches of 19951996, and 1997, winning a half blue.

He was educated at Christ’s College, Finchley, a state-funded comprehensive school in East Finchley, North London. He completed his GCE Advanced Level and Scholarship Level exams early at the age of 15 and 16.

Demis began a career in video games at British studio Bullfrog Productions, co-designing and lead programming on the classic game Theme Park at 17 years old, alongside the legendary games designer Peter Molyneux.

Released in 1994, Theme Park sold several million copies and won a Golden Joystick award and inspired a whole genre of management sim games. The game set players the task of building a successful theme park in the UK with just a few thousand pounds and a small plot of land.

It went into great detail with planning and strategies, which for success in the game included putting more salt in crisps so people would buy more drinks, making queues long but fast-moving and obscuring the destination, hiring a cleaner to clean the toilets, and placing the exit of one roller-coaster close to the entrance of the next to minimize the walking distance between attractions.

Demis left Bullfrog to study Computer Science at Queens’ College, the University of Cambridge, which in 1953 had the world’s first undergraduate computer science course. He graduated from Queens’ College Cambridge with a double first-class honours degree in 1997.

After graduating from Cambridge, Demis worked at Lionhead Studios where he yet again joined renowned games designer Peter Molyneux who had recently founded the company. At Lionhead, Demis worked as lead AI programmer on the 2001 “god” game Black & White.

In 1998, he left Lionhead Studios and founded his own London-based independent games developer, Elixir Studios. The company grew to 60 people strong and signed deals with large publishers Eidos InteractiveVivendi Universal, and Microsoft. His games included Republic: The Revolution and Evil Genius, which were both BAFTA-nominated.

The release of Elixir’s first game, Republic: The Revolution, a highly ambitious and unusual political simulation game, was delayed due to its huge scope. The final game was reduced from its original vision and greeted with lukewarm reviews, receiving a Metacritic score of 62/100. The game was “way ahead of the market”.

“Now it still might not be a good idea — just because you passionately believe in it, that doesn’t make it a good idea.”

“You don’t want to be 50 years ahead of your time, you want to be five years ahead. And with some of the things we were trying to do in my first [Elixir] company, I think we were way ahead of the hardware and frankly where the market was,” Demis said.

The intellectual property and technology rights from Elixir Studios were sold to various games publishers and the studio was closed in April 2005.

In 1999, aged 23, he won the Mind Sports Olympiad — an annual international multi-disciplined competition for games of mental skill. He won it a record five times before retiring from competitive play in 2003.

Demis was later elected as a fellow of the Royal Society of Arts in 2009 for his influential game-design work and contribution to the games industry.

Following Elixir Studios, Demis returned to academia and a career in cognitive neuroscience, to allow him to return to his primary interest in artificial intelligence. He earned his Ph.D. in Cognitive Neuroscience from University College London (UCL) in 2009 supervised by Eleanor Maguire with the thesis Neural processes underpinning episodic memory. He sought to find inspiration in the human brain for new AI algorithms.

He continued his neuroscience and artificial intelligence research as a visiting scientist jointly at Massachusetts Institute of Technology (MIT) and Harvard University, before earning a Henry Wellcome postdoctoral research fellowship to the Gatsby Charitable Foundation Computational Neuroscience Unit, UCL in 2009.

Working in the field of autobiographical memory and amnesia, he co-authored several influential papers published in NatureScienceNeuron, and PNAS. One of his most highly cited papers, published in the Proceedings of the National Academy of Sciences of the United States of America, showed systematically for the first time that patients with damage to their hippocampus, known to cause amnesia, were also unable to imagine themselves in new experiences.

The finding established a link between the constructive process of imagination and the reconstructive process of episodic memory recall. Based on this work and a follow-up functional magnetic resonance imaging (fMRI) study, Demis developed a new theoretical account of the episodic memory system identifying scene construction, the generation, and online maintenance of a complex and coherent scene, as a key process underlying both memory recall and imagination. This work received widespread coverage in the mainstream media (The New York Times, 2007) and was listed in the top 10 scientific breakthroughs of the year in any field by the journal Science.

In 2010, Demis co-founded DeepMind, a London-based machine learning AI startup, with Shane Legg and Mustafa Suleyman. Demis and Suleyman had been friends since childhood, and he met Legg when both were postdocs at University College London’s Gatsby Computational Neuroscience Unit. Demis also recruited his university friend and Elixir partner David Silver.

DeepMind’s mission is to “solve intelligence” and then use intelligence “to solve everything else”. More concretely, DeepMind aims to meld insights from neuroscience and machine learning with new developments in computing hardware to unlock increasingly powerful general-purpose learning algorithms that will work towards the creation of an artificial general intelligence (AGI).

The company has focused on training learning algorithms to master games, and in December 2013 it famously announced that it had made a pioneering breakthrough by training an algorithm called a Deep Q-Network (DQN) to play Atari games at a superhuman level by only using the raw pixels on the screen as inputs.

DeepMind’s early investors included several high-profile tech entrepreneurs. On 27 January in 2014, Google purchased DeepMind for £400 million, although it has remained an independent entity based in London.

Since the Google acquisition, the company has made a number of significant achievements, among others their programs have learned to diagnose eye diseases as effectively as the world’s top doctors, to save 30% of the energy used to keep data centers cool, and to predict the complex 3D shapes of proteins with AlphaFold — which could one day transform how drugs are invented.

The most notable was perhaps the creation of AlphaZero which achieved within 24 hours a superhuman level of play in the games of chess and shogi (Japanese chess) as well as Go, and convincingly defeated a world-champion program in each case, including renowned AlphaGo, a program that defeated world champion Lee Sedol and winning 4–1 at the complex game of Go in March 2016. Go had been considered a holy grail of AI, for its high number of possible board positions and resistance to existing programming techniques.

Their most recent groundbreaking use of AlphaZero, is its AlphaStar algorithm which came out in January 2019 and is the first AI to beat a professional player at StarCraft II. StarCraft is a game that has been around since 1998 and is the 5th best selling PC game of all time. It is considered one of the most complex Real-Time Strategy (RTS) games and it is also one of the longest played RTS games. This makes it a great game to test AI against for more real world type challenges.

Additional DeepMind accomplishments include creating a Neural Turing machine, advancing research on AI safety with their Safely Interruptible Agents, and the creation of a partnership with the National Health Service (NHS) of the United Kingdom and Moorfields Eye Hospital to improve medical service and identify the onset of degenerative eye conditions.

More recently, DeepMind turned its artificial intelligence to protein-folding, one of the toughest problems in science. In December 2018, DeepMind’s tool AlphaFold won the 13th Critical Assessment of Techniques for Protein Structure Prediction (CASP) by successfully predicting the most accurate structure for 25 out of 43 proteins.

“This is a lighthouse project, our first major investment in terms of people and resources into a fundamental, very important, real-world scientific problem,” Demis said to the Guardian.

DeepMind has also been responsible for technical advancements in machine learning, having produced a number of award-winning papers. In particular, the company has made significant advances in deep learning and reinforcement learning, and pioneered the field of deep reinforcement learning which combines these two methods. Demis has predicted that Artificial Intelligence will be “one of the most beneficial technologies of mankind ever” but that significant ethical issues remain.

Their significant breakthroughs are as follows: AlphaZero (pub), AlphaGo (pub1pub2), Deep Q-Network (DQN) (pub1pub2), A neural network with dynamic memory (pub), WaveNet (A generative model for raw audio) (pub), and The Generative Query Network (GQN), a Neural scene representation and rendering (pub).

In one of his interviews Three truths about AI at TechRepublic, Demis said:

“I would actually be very pessimistic about the world if something like AI wasn’t coming down the road.”.

“The reason I say that is that if you look at the challenges that confront society: climate change, sustainability, mass inequality — which is getting worse — diseases, and healthcare, we’re not making progress anywhere near fast enough in any of these areas.

“Either we need an exponential improvement in human behavior — less selfishness, less short-termism, more collaboration, more generosity — or we need an exponential improvement in technology.

“If you look at current geopolitics, I don’t think we’re going to be getting an exponential improvement in human behavior any time soon.

“That’s why we need a quantum leap in technology like AI.”

Demis and DeepMind are keeping their developed technologies in their own hands and control from Google, with a signed contract called the Ethics and Safety Review Agreement. It puts control of DeepMind’s core AGI technology, whenever it may be created, in the hands of a governing panel known as the Ethics Board.

Far from being a cosmetic concession from Google, the Ethics Board gives DeepMind solid legal backing to keep control of its most valuable and potentially most dangerous technology, according to the same source.

DeepMind is also investing into Safety and Ethics. They published the Safely Interruptible Agents paper which describes the development of kill switch for AI or how to implement a theoretical Big Red Button. AI can provide great benefits, but like all technology, it can have negative impacts unless it’s built and used responsibly. AI systems can only benefit the world if they are made reliable and safe. At DeepMind, they support open research and investigation into the wider impacts of AI, by making all the research public and available, and working on technical safety, ethics, and public engagement.

Demis is a five-time World Pentamind Champion and two-time World Decamentathlon Champion, recipient of the Royal Society’s Mullard Award, the Silver Medal from the Royal Academy of Engineering, the American Academy of Achievement’s Golden Plate Awardthe Honorary degree by Imperial College London, the Outstanding Achievement in Science and Technology (2017) at The Asian Awards, and became a Fellow of the Royal Academy of Engineering, a Fellow of the Royal Society of Arts (FRSA) in 2009, a Fellow Benefactor at Queens’ College in Cambridge, an Honorary Fellow at University College London, elected a Fellow of the Royal Society (FRS) in May 2018, and finally appointed Commander of the Order of the British Empire (CBE) in the 2018 New Year Honours for “services to Science and Technology”.

Among received awards, Demis has been also chosen by Science Magazine for Top 10 Scientific Breakthroughs of 2007 (for neuroscience research on imagination) and for AlphaGo among Top 10 Scientific Breakthroughs of 2016, as third most influential Londoner in 2014 and 6th most influential Londoner in 2016 by the London Evening Standard, top 50 Entrepreneurs in Europe (2015) and Digital Entrepreneur of the Year (2016) by Financial Times, selected by Nature’s Ten people who matter this year (2016), selected among Time 100’s The 100 Most Influential People (2017), and listed on Wired’s ‘Smart 50’ (2013).

Demis also cashed at the World Series of Poker six times including in the Main Event, and became Diplomacy World Team Champion in 2004, 4th in 2006 World Championship, and 3rd in the 2004 European Championship.

He is a member of The American Academy of Achievement, being inducted in 2017, the Centre for the Future of Intelligence,

Read DeepMind and Google: the battle to control artificial intelligence as appeared in the Economist’s 1843 Stories of an Extraordinary World.

Read ZDNet’s news on Demis’ talk at the Institute for Advance Study in Princeton, New Jersey about where AI has been and where it’s going.

Read Google’s StarCraft II victory shows AI improves via diversity, invention, not reflexes at ZDNet.

Read Business Insider article The incredible life of DeepMind founder Demis Hassabis, the computer whiz who sold his AI lab to Google for £400 million.

Read the article with Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality.

Read some of the interviews with Demis following the acquisition by Google with Evening Standard (2014)Telegraph (2014)The Guardian (2014)Financial Times (2015), and PCGamesN (2016).

Read the Wired Featured article DeepMind: inside Google’s super-brain.

Listen to interview, the Desert Island Discs series and podcast The Disrupters, made with Demis at BBC.

Watch Demis’ lectures and talks: Creativity and AI – The Rothschild Foundation Lecture (2018)The Future of AI and Science: Cheltenham Science Festival 2018The Power of Self-Learning Systems – Institute of Advanced Study (2019), Falling Walls’ How Deep Learning Can Give Birth to General Artificial Intelligence, and Talking about Transfer Learning is key to AGI.

Read his co-authored papers Mastering the game of Go with deep neural networks and tree search (2016) and Human-level control through deep reinforcement learning (2015).

Visit his Personal Homepage, the Academy of AchievementThe Royal Society ProfileGoogle Scholar profileLinkedIn profileStanford page, and Fide Chess Profile. Follow him on FacebookIMDB, and Twitter. Follow DeepMind’s Blog and Demis’ Wikipedia page.