What Is Google’s DeepMind?

September 16, 2016 • Devin Partida


DeepMind, an artificial intelligence company that Google acquired in 2014, creates neural networks that learn in ways similar to the way humans learn. In the future, this technology could have an impact on everything from the way you use your smartphone to your next doctor’s visit.

What Is DeepMind?

Shane Legg, Mustafa Suleyman and Demis Hassabis founded DeepMind in London in 2010. Google purchased DeepMind in January 2014 for a reported $400 million, leaving Hassabis in charge as CEO.

The Google subsidiary researches and develops Artificial Intelligence (AI), specifically a branch of AI known as machine learning. Instead of relying on specific input from programmers for a given situation, DeepMind uses general learning algorithms to analyze huge sets of data and forecast outcomes. These algorithms use strategies such as deep neural networks and reinforcement learning.

DeepMind made headlines in March 2016 when its computer program, AlphaGo, beat Lee Sodel, one of the world’s top-ranked Go players. Go, a very complex ancient Chinese game, has been regarded as something of an ultimate challenge for artificial intelligence.

What makes AlphaGo’s victory especially important is that the program used machine learning to figure the game out. It wasn’t specifically programmed to play the game. It learned how to play on its own.

DeepMind technology has used similar learning techniques to master various video games. But there’s more to DeepMind’s goals than just beating games. Every game DeepMind’s technology masters has some relevance to its ultimate plan — focusing on real-world problems and improving human decision-making.

How Might DeepMind Help You in the future?

DeepMind, and AI in general, will likely have extensive effects on our futures. Alphabet Chairman, Eric Schmidt, says he believes Alphabet, Google’s parent company, could use DeepMind to improve every single one of their products.

AI could be useful in almost any situation in which lots of data needs to be used to predict a result — and potentially make you more productive. Here are just some of the ways DeepMind could help you in the future.


It Could Make Your Search Results More Personalized

When you search for something on Google’s search engine, Google sifts through tons of data trying to find you the results you’re looking for. Its likely that DeepMind will improve these services by learning your individual preferences and, therefore, showing you more accurately the results you want.

It Could Make Your Smartphone Assistant More Helpful

Presently, most smartphone assistants, such as Apple’s Siri and Microsoft’s Cortana, follow pre-programmed paths, which limits their usefulness. DeepMind wants to use deep learning to improve these assistants, so that they have a better understanding of what the user wants to do. Hassabis even imagines a smartphone assistant that could plan and book an entire vacation for you based on your preferences.

DeepMind has also made breakthroughs in artificial speech. They created WaveNet, a system that learns from audio it’s given and creates its own sounds, rather than rearranging snippets of prerecorded human speech. DeepMind was able to use WaveNet for advanced speech recognition. These advances could it make it much easier to communicate with your smartphone assistant in the future.

It Could Help Doctors Spot Health Issues

Alphabet will likely use AI in Verily, a division of the company that focuses on healthcare.

DeepMind has a partnership with the UK’s National Health Service. Through this partnership, DeepMind is researching how to use its technology to more accurately discover eye health issues by analyzing the digital scans currently used by doctors. DeepMind hopes to make healthcare more personalized by helping doctors determine what is best for a specific patients based on that patient’s health history.

It Could Save You Money on Your Bills

DeepMind proved to be worth Google’s money when it helped slash the costs of running their data centers. Through controlling 120 different variables and using the same learning techniques it used to master the game of Go, DeepMind improved power usage efficiency by 15%.

That same technology could one day be used to make your home more energy efficient. That’s good news for your wallet and for the environment.

Technology inserts itself into our lives gradually. Each advancement is like a lightyear — especially as consumers seek convenience and prioritize finding easier ways to do things to get more done each day.

You might not notice how AI like DeepMind influences your day-to-day life now, but soon these advancements will be daily, commonplace experiences.