London-based brain-tech company DeepMind was founded in 2010 and acquired by Google in 2014 for more than $500 million. Right now, DeepMind has over 400 research engineers and scientists, more than 250 of them with PhDs, making this pool of human resources arguably the most formidable collection of scientific brain power focused on a single subject: artificial intelligence. DeepMind is also a fusion of the best scientific thinking of academia with with the speed and focus of the best Silicon Valley startups.
The company’s mission involves a two-step process. First, they intend to solve intelligence by understanding natural intelligence, then recreating it artificially. Second, they intend to use AI to solve everything else.
As Demis Hassabis, cofounder and CEO of Google DeepMind explained in an article he wrote for the Financial Times, the approach that DeepMind takes focuses on ‘learning and generality’, the objective of which is to develop the kind of AI that’s needed for science. He says, “If we want computers to discover new knowledge, then we must give them the ability to truly learn for themselves.”
To achieve their mission, they plan to build the world’s first general purpose learning system, one that learns everything from data. He says that the algorithms they are working on learn from experience, acquiring knowledge based on ‘some form of sensory reality rather than in abstract symbols’. Then they require the learning to be general, meaning, the same system using the same parameters should be able to perform well across a diverse range of tasks, across different environments, being able to transfer their knowledge seamlessly from one domain to another.
They start off with games as their testing ground, then translate the algorithms so they can be applied in real-life situations. More importantly, they use neuroscience as their inspiration for coming up with new concepts and ideas, because ‘the brain is the only existence proof we have that a general-purpose experience-based learning system is even possible.’
Contrary to what many perceive AI as being a threat to humanity, Hassabis believes that ‘humans and machines collaborating in a complimentary way will allow us to achieve really amazing things’. As an incredibly powerful technology, the purpose of AI is not to surpass humanity’s capability, but to ‘augment our human ingenuity and unlock our true potential’.
Comparing AI with the Hubble Telescope, he asserts that AI is the ultimate tool to explore the universe, providing support to all kinds of experts, by being able to identify patterns and sources that the human eyes cannot perceive on their own.
And because for the well-informed there is no doubt about AI’s transformational potential in terms of changing everything about our world, the company is well aware that they should use this human-replacing technology ‘responsibly, ethically, and for everyone’s benefit’.
With these positive thoughts in mind, consider watching this video to find out more about DeepMind’s plans.