Google's secretive artificial intelligence researchers have revealed a computer that they hope will one day be able to program itself. Developers at Google's secretive DeepMind start-up, which it bought for $400 million earlier this year, are attempting to mimic some of the properties of the human brain's short-term working memory. By combining the way ordinary computers work with the way the human brain works, the researchers hope the machine will learn to program itself. Described as a 'Neural Turing Machine', it can learns as it stores memories, and later retrieve them to perform logical tasks beyond those it has been trained to do.'We have introduced the Neural Turing Machine, a neural network architecture that takes inspiration from both models of biological working memory and the design of digital computers,' the researchers wrote. 'Our experiments demonstrate that it is capable of learning simple algorithms from example data and of using these algorithms to generalise well outside its training regime.'The new computer is effectively a type of neural network that has been adapted to work with an external memory. The result is a computer that learns as it stores memories and can later retrieve them to perform logical tasks beyond those it has been trained to do. 'We extend the capabilities of neural networks by coupling them to external memory resources, which they can interact with by attentional processes, the team wrote.
'The combined system is analogous to a Turing Machine or Von Neumann architecture but is differentiable end-toend, allowing it to be efficiently trained with gradient descent. 'Preliminary results demonstrate that Neural Turing Machines can infer simple algorithms such as copying, sorting, and associative recall from input and output examples.' Last week the California tech giant revealed it has has teamed up with two of Oxford University's artificial intelligence (AI) teams to help machines better understand users, and improve visual recognition systems using deep learning.This partnership follows reports Google is also developing superfast 'quantum' chips modelled on the human brain, to make searches and software more intuitive.'It is a really exciting time for AI research these days, and progress is being made on many fronts including image recognition and natural language understanding,' wrote Demis Hassabis, co-founder of DeepMind and vice president of engineering at Google in a blog post. 'We are delighted to announce a partnership with Oxford University to accelerate Google's research efforts in these areas.' Google DeepMind will be working with two of Oxford's AI research teams.Professor Nando de Freitas, Professor Phil Blunsom, Dr Edward Grefenstette and Dr Karl Moritz Hermann, who teamed up earlier this year to co-found Dark Blue Labs, will be leading research to help machines better understand what users are saying to them. Also joining the DeepMind team will be Dr Karen Simonyan, Max Jaderberg and Professor Andrew Zisserman, experts in computer vision systems. As co-founders of Vision Factory, their aim is to improve visual recognition systems using deep learning. Dr Simonyan and Professor Zisserman developed one of the winning systems at the recent 2014 ImageNet competition. Google DeepMind has hired all seven founders of these startups with the three professors holding joint appointments at Oxford University where they will continue to spend part of their time.'These exciting partnerships underline how committed Google DeepMind is to supporting the development of UK academia and the growth of strong scientific research labs,' continued Mr Hassabis. As a part of the collaboration, Google DeepMind will be donating to establish a research partnership with the Computer Science Department and the Engineering Department at Oxford University. 'We are thrilled to welcome these extremely talented machine learning researchers to the Google DeepMind team and are excited about the potential impact of the advances their research will bring,' concluded Mr Hassabis. Google is also reportedly working on a super-fast 'quantum' computer chip as part its vision to one day have machines think like humans.
The California-based group has teamed up with leading physicist John Martinis to build processors based on quantum theories. The new hire is part of a 'hardware initiative' to design and build chips operating on sub-atomic levels in ways that makes them much faster than existing processors. Standard computers deal with binary data expressed in zeroes and ones. However, quantum computing uses the behaviour of sub-atomic particles to encode data. Experts believe that a quantum bit, which can have two states at the same time, may be able hugely improve the speed and power of computing. Professor Martinis currently works at University of California, Santa Barbara, and is one of the most prolific researchers in the area of artificial intelligence. But Google is also aware of the dangers involved with AI and machine learning. So much so that in January it set up an ethics board to oversee its work in these fields. The ethics board is to ensure the projects are not abused. The DeepMind-Google ethics board is set to create a series of rules and restrictions over the use of the technology.
AI IS 'MORE DANGEROUS THAN NUKES', CLAIMS ELON MUSKE
lon Musk is one of the driving forces behind super-intelligent computers that could improve everything from space travel to electric cars. But the Tesla-founder claims the technology could someday be more harmful than nuclear weapons. Last month, the billionaire tweeted a recommendation for a book that looks at a robot uprising, claiming 'We need to be super careful with AI. Potentially more dangerous than nukes.' Musk referred to the book 'Superintelligence: Paths, Dangers, Strategies', a work by Nick Bostrom that asks major questions about how humanity will cope with super-intelligent computers. Mr Bostrom has also argued that the world is fake and we are living in a computer simulation. In a later tweet, Musk wrote: 'Hope we're not just the biological boot loader for digital superintelligence. Unfortunately, that is increasingly probable.' Musk's tweets follows a similar comment in June in which the Tesla-founder said believes that a horrific 'Terminator-like' scenario could be created from research into artificial intelligence. The 42-year-old is so worried, he is investing in AI companies, not to make money, but to keep an eye on the technology in case it gets out of hand.
The California-based group has teamed up with leading physicist John Martinis to build processors based on quantum theories. The new hire is part of a 'hardware initiative' to design and build chips operating on sub-atomic levels in ways that makes them much faster than existing processors. Standard computers deal with binary data expressed in zeroes and ones. However, quantum computing uses the behaviour of sub-atomic particles to encode data. Experts believe that a quantum bit, which can have two states at the same time, may be able hugely improve the speed and power of computing. Professor Martinis currently works at University of California, Santa Barbara, and is one of the most prolific researchers in the area of artificial intelligence. But Google is also aware of the dangers involved with AI and machine learning. So much so that in January it set up an ethics board to oversee its work in these fields. The ethics board is to ensure the projects are not abused. The DeepMind-Google ethics board is set to create a series of rules and restrictions over the use of the technology.