AI Self-Training Goes Forward at Google DeepMind

By Doug Black

October 19, 2017

Imagine if all the atoms in the universe could be added up into a single number. Big number, right? Maybe the biggest number conceivable. But wait, there’s a bigger number out there. We’re told that Go, the world’s oldest board game, has more possible board positions than there are atoms in the universe. Urban myth? All right, let’s say Go has half as many positions as there are atoms. Make it a tenth. The point is: Go complexity is beyond measure.

DeepMind, Google’s AI research organization, announced today in a blog that AlphaGo Zero, the latest evolution of AlphaGo (the first computer program to defeat a Go world champion) trained itself within three days to play Go at a superhuman level (i.e., better than any human) – and to beat the old version of AlphaGo – without leveraging human expertise, data or training.

The absence of human training may have “liberated” AlphaGo Zero to find new ways to play Go that humans don’t know, putting the new system beyond the talents of the human-trained AlphaGo.

Richard Windsor, analyst at Edison Investment Research, London, notes that today’s announcement is an important step forward on one of the three big AI challenges which, he said, are:

  • AI systems that can be trained with less data
  • AI that takes lessons learned from one task and applies it across multiple tasks
  • AI that builds its own models

“DeepMind has been able to build a new Go (AlphaGo Zero) algorithm that relies solely on self-play to improve and within 36 hours was able to defeat AlphaGo Lee (the one that beat [professional Go player] Lee Sedol) 100 games to 0…,” Windsor said. “DeepMind’s achievement represents a huge step forward in addressing the first challenge as AlphaGo Zero used no data at all…”

According to DeepMind, previous versions of AlphaGo were trained on the basis of thousands of human games. But AlphaGo Zero “skips this step and learns to play simply by playing games against itself, starting from completely random play.” In doing so, it quickly surpassed human level of play and went undefeated against AlphaGo.

The new self-training algorithm, according to the DeepMind blog, is significant for AI systems to take on problems for which “human knowledge may be too expensive, too unreliable or simply unavailable. As a result, a long-standing ambition of AI research is to bypass this step, creating algorithms that achieve superhuman performance in the most challenging domains with no human input.”

DeepMind said AlphaGo Zero uses a novel form of reinforcement learning in which the system starts off with a neural network that knows nothing about Go. “It then plays games against itself, by combining this neural network with a powerful search algorithm. As it plays, the neural network is tuned and updated to predict moves, as well as the eventual winner of the games.”

AhaGo has become progressively more efficient thanks to hardware gains and more recently algorithmic advances (Source: DeepMind)

The updated neural network is then recombined with the search algorithm to create a new, stronger version of AlphaGo Zero, and the process begins again, improving incrementally with each game. (The algorithmic change also significantly improves system efficiency, see graphic at right.)

“This technique is more powerful than previous versions of AlphaGo because it is no longer constrained by the limits of human knowledge. Instead, it is able to learn tabula rasa from the strongest player in the world: AlphaGo itself,” said DeepMind.

Put another way by Windsor: “It is almost as if the use of human data limited the potential of the machine’s ability to maximize its potential.”

While the new system makes strides against the self-training Big AI Challenge, Windsor expressed doubts that it addresses the third challenge (automated model building) because it used a model already used by the previous version of AlphaGo.

“…the system of board assessment and move prediction (but not the experience) used in AlphaGo Lee was also built into AlphaGo Zero,” said Windsor. “Hence, we think that this system was instead using a framework that had already been developed to play and applying reinforcement learning to improve, rather than building its own models.”

But this isn’t to minimize the achievement of AlphaGo Zero, nor to quell those (such as Elon Musk) who worry that human intelligence will eventually be dwarfed by AI, with potential dystopic implications.

“What will really have the likes of Elon Musk quaking in their boots is the fact that AlphaGo Zero was able to obtain a level of expertise of Go that has never been achieved by a human mind,” Windsor said.

Having said that, include Windsor among those who don’t believe machines will enslave the human race. He also said that DeepMind may have trouble applying its achievement elsewhere.

“Many of the other digital ecosystems have been trying to use computer generated images to train image and video recognition algorithms but there has been no real success to date and we suspect that taking what DeepMind has achieved and applying it to real world AI problems like image and video recognition will be very difficult,” he said, explaining that “the Go problem is based on highly structured data in a clearly defined environment whereas images, video, text, speech and so on are completely unstructured.”

But DeepMind sounded a more optimistic note on the broader applicability of AlphaGo Zero teaching itself new and incredibly complicated tricks.

“These moments of creativity give us confidence that AI will be a multiplier for human ingenuity, helping us with our mission to solve some of the most important challenges humanity is facing…. If similar techniques can be applied to other structured problems, such as protein folding, reducing energy consumption or searching for revolutionary new materials, the resulting breakthroughs have the potential to positively impact society.”

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

SCA23: Pawsey’s Mark Stickells on Sustainable Australian Supercomputing

March 17, 2023

“While the need for supercomputing is great, we have, in my view, reached a tipping point,” said Mark Stickells, executive director of Australia’s Pawsey Supercomputing Centre, as he opened his keynote (“Energy E Read more…

Optical I/O Technology Needed for Zettascale, Say Top Chipmakers

March 16, 2023

Optical I/O is being singled out by top companies to push computing beyond exascale and into zettascale. The technology was singled out in a recent speech by AMD CEO Lisa Su as a critical technology to reach zettascale c Read more…

Tasty CHIPS – New MEC Program to Expand US Prototyping Capabilities Gains Steam

March 16, 2023

Sometime later this year, perhaps around July, the Department of Defense is expected to announce the sites and focus of up to nine hubs associated with the Microelectronics Commons (MEC) program. Funded and broadly descr Read more…

2023 Winter Classic: Mentor Interview, HPE

March 14, 2023

In our most recent update, “Triumph and Tragedy with HPL/HPCG”, we detailed how our dozen 2023 Winter Classic Invitational cluster competition teams dealt with their Linpack/HPCG module, mentored by HPE. In this episode of our incredibly popular 2023 Winter Classic Studio Update Show, we... Read more…

Leibniz QIC’s Mission to Coax Qubits and Bits to Work Together

March 14, 2023

Four years after passing the U.S. National Quantum Initiative Act and decades after early quantum development and commercialization efforts started – think D-Wave Systems and IBM, for example – the U.S. quantum lands Read more…

AWS Solution Channel

Shutterstock 1679096101

Building a 4x faster and more scalable algorithm using AWS Batch for Amazon Logistics

Amazon Logistics’ science team created an algorithm to improve the efficiency of their supply-chain by improving planning decisions. Initially the algorithm was implemented in a sequential way using a monolithic architecture executed on a single high performance computational node on AWS Cloud. Read more…

 

Get the latest on AI innovation at NVIDIA GTC

Join Microsoft at NVIDIA GTC, a free online global technology conference, March 20 – 23 to learn how organizations of any size can power AI innovation with purpose-built cloud infrastructure from Microsoft. Read more…

Pawsey Supercomputing Targets Detailed Regional Climate Projections

March 13, 2023

The Pawsey Supercomputing Centre in Australia is putting its shiny new Setonix supercomputer (ranked fourth on the most recent Top500 list) to work on an important climate change research project. The project, led by Jat Read more…

SCA23: Pawsey’s Mark Stickells on Sustainable Australian Supercomputing

March 17, 2023

“While the need for supercomputing is great, we have, in my view, reached a tipping point,” said Mark Stickells, executive director of Australia’s Pawsey Read more…

Optical I/O Technology Needed for Zettascale, Say Top Chipmakers

March 16, 2023

Optical I/O is being singled out by top companies to push computing beyond exascale and into zettascale. The technology was singled out in a recent speech by AM Read more…

Tasty CHIPS – New MEC Program to Expand US Prototyping Capabilities Gains Steam

March 16, 2023

Sometime later this year, perhaps around July, the Department of Defense is expected to announce the sites and focus of up to nine hubs associated with the Micr Read more…

Leibniz QIC’s Mission to Coax Qubits and Bits to Work Together

March 14, 2023

Four years after passing the U.S. National Quantum Initiative Act and decades after early quantum development and commercialization efforts started – think D- Read more…

Intel Hopes to Stop Server Beating from AMD Next Year

March 13, 2023

After getting bruised in servers by AMD, Intel hopes to stop the bleeding in the server market with next year's chip offerings. The difference-making products will be Sierra Forest and Granite Rapids, which are due out in 2024, said Dave Zinsner, chief financial officer at Intel, last week at the Morgan Stanley Technology, Media and Telecom conference. Read more…

White House Budget Request Includes Funding for Leadership-Class Computing Facility

March 10, 2023

The U.S. government is dedicating a record amount of $25 billion as part of the 2024 budget to emerging technologies as the country looks to counter the technology threat from China. The budget includes billions of dollars earmarked to boost the supercomputing infrastructure, semiconductors, and cutting-edge technologies such as artificial intelligence and quantum computing. The technology... Read more…

Inside NCSA’s Nightingale Cluster, Designed for Sensitive Data

March 10, 2023

The emergence of Covid in 2020 saw an explosion in HPC-powered health research. As the pandemic raged on, though, one limiting factor became increasingly clear: Read more…

Top HPC Players: It’s Time to Get Serious About Security

March 9, 2023

Time’s up: nearly everyone agrees it’s about time to become serious about bringing security safeguards to high-performance computing systems, which has been Read more…

CORNELL I-WAY DEMONSTRATION PITS PARASITE AGAINST VICTIM

October 6, 1995

Ithaca, NY --Visitors to this year's Supercomputing '95 (SC'95) conference will witness a life-and-death struggle between parasite and victim, using virtual Read more…

SGI POWERS VIRTUAL OPERATING ROOM USED IN SURGEON TRAINING

October 6, 1995

Surgery simulations to date have largely been created through the development of dedicated applications requiring considerable programming and computer graphi Read more…

U.S. Will Relax Export Restrictions on Supercomputers

October 6, 1995

New York, NY -- U.S. President Bill Clinton has announced that he will definitely relax restrictions on exports of high-performance computers, giving a boost Read more…

Dutch HPC Center Will Have 20 GFlop, 76-Node SP2 Online by 1996

October 6, 1995

Amsterdam, the Netherlands -- SARA, (Stichting Academisch Rekencentrum Amsterdam), Academic Computing Services of Amsterdam recently announced that it has pur Read more…

Cray Delivers J916 Compact Supercomputer to Solvay Chemical

October 6, 1995

Eagan, Minn. -- Cray Research Inc. has delivered a Cray J916 low-cost compact supercomputer and Cray's UniChem client/server computational chemistry software Read more…

NEC Laboratory Reviews First Year of Cooperative Projects

October 6, 1995

Sankt Augustin, Germany -- NEC C&C (Computers and Communication) Research Laboratory at the GMD Technopark has wrapped up its first year of operation. Read more…

Sun and Sybase Say SQL Server 11 Benchmarks at 4544.60 tpmC

October 6, 1995

Mountain View, Calif. -- Sun Microsystems, Inc. and Sybase, Inc. recently announced the first benchmark results for SQL Server 11. The result represents a n Read more…

New Study Says Parallel Processing Market Will Reach $14B in 1999

October 6, 1995

Mountain View, Calif. -- A study by the Palo Alto Management Group (PAMG) indicates the market for parallel processing systems will increase at more than 4 Read more…

Leading Solution Providers

Contributors

CORNELL I-WAY DEMONSTRATION PITS PARASITE AGAINST VICTIM

October 6, 1995

Ithaca, NY --Visitors to this year's Supercomputing '95 (SC'95) conference will witness a life-and-death struggle between parasite and victim, using virtual Read more…

SGI POWERS VIRTUAL OPERATING ROOM USED IN SURGEON TRAINING

October 6, 1995

Surgery simulations to date have largely been created through the development of dedicated applications requiring considerable programming and computer graphi Read more…

U.S. Will Relax Export Restrictions on Supercomputers

October 6, 1995

New York, NY -- U.S. President Bill Clinton has announced that he will definitely relax restrictions on exports of high-performance computers, giving a boost Read more…

Dutch HPC Center Will Have 20 GFlop, 76-Node SP2 Online by 1996

October 6, 1995

Amsterdam, the Netherlands -- SARA, (Stichting Academisch Rekencentrum Amsterdam), Academic Computing Services of Amsterdam recently announced that it has pur Read more…

Cray Delivers J916 Compact Supercomputer to Solvay Chemical

October 6, 1995

Eagan, Minn. -- Cray Research Inc. has delivered a Cray J916 low-cost compact supercomputer and Cray's UniChem client/server computational chemistry software Read more…

NEC Laboratory Reviews First Year of Cooperative Projects

October 6, 1995

Sankt Augustin, Germany -- NEC C&C (Computers and Communication) Research Laboratory at the GMD Technopark has wrapped up its first year of operation. Read more…

Sun and Sybase Say SQL Server 11 Benchmarks at 4544.60 tpmC

October 6, 1995

Mountain View, Calif. -- Sun Microsystems, Inc. and Sybase, Inc. recently announced the first benchmark results for SQL Server 11. The result represents a n Read more…

New Study Says Parallel Processing Market Will Reach $14B in 1999

October 6, 1995

Mountain View, Calif. -- A study by the Palo Alto Management Group (PAMG) indicates the market for parallel processing systems will increase at more than 4 Read more…

SC22 Booth Videos

AMD @ SC22
Altair @ SC22
AWS @ SC22
Ayar Labs @ SC22
CoolIT @ SC22
Cornelis Networks @ SC22
DDN @ SC22
Dell Technologies @ SC22
HPE @ SC22
Intel @ SC22
Intelligent Light @ SC22
Lancium @ SC22
Lenovo @ SC22
Microsoft and NVIDIA @ SC22
One Stop Systems @ SC22
Penguin Solutions @ SC22
QCT @ SC22
Supermicro @ SC22
Tuxera @ SC22
Tyan Computer @ SC22
  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire