Google has applied software developed by London research hub DeepMind to its virtual assistant in a move that makes the £400 million acquisition of DeepMind that bit more worthwhile.
Since being acquired in 2014, DeepMind has carried out a vast amount of research in the field of artificial intelligence but most of it can’t be directly applied to Google’s products.
It is best-known for developing AI systems that have mastered games like Space Invaders and Go. However, Google has been slow to integrate DeepMind’s technology into its products, with just one data centre efficiency project announced so far, albeit on a global scale.
In a bid to increase collaboration between Google and DeepMind, a new “DeepMind for Google” team was set up and DeepMind also opened a small office at Google’s headquarters in Mountain View, California.
Now the company’s WaveNet neural network is being used to generate the Google Assistant voices for US English and Japanese.
DeepMind announced the update in a blog post on its website on Wednesday that was co-authored by DeepMind research scientist Aäron van den Oord, research scientist Tom Walters, and Google Speech software engineer Trevor Strohman.
“Just over a year ago we presented WaveNet, a new deep neural network for generating raw audio waveforms that is capable of producing better and more realistic-sounding speech than existing techniques. At that time, the model was a research prototype and was too computationally intensive to work in consumer products.
“But over the last 12 months we have worked hard to significantly improve both the speed and quality of our model and today we are proud to announce that an updated version of WaveNet is being used to generate the Google Assistant voices for US English and Japanese across all platforms.”
DeepMind’s AI research cost £164 million last year
DeepMind’s work is costing Google parent company Alphabet tens of millions of pounds, with most of that going on hiring top people.
The company spent £104.8 million on “staff costs and other related costs” last year, according to a document filed with Companies House this week.
A DeepMind spokesperson said: “We’re really proud that some of the world’s most exciting AI research and real-world application is taking place right here in London. We intend to keep investing in our scientific mission, and to work with the world’s brightest minds to tackle society’s most complex problems.”
Overall, DeepMind made a loss of £164 million in total in 2016, a significant increase on the £54 million loss it posted in 2015.
However, the company, brought in revenue for the first time in 2016, posting a turnover of £40 million. The turnover relates to the DeepMind’s projects with Google.