New hardware launches from Alphabet Inc\u2019s Google on Wednesday showed how the acquisition of London-based artificial intelligence company DeepMind might start to generate revenue rather than just research papers.<\/p>\n
Alphabet bought DeepMind for a reported $400 million in 2014. The firm has produced a steady stream of machine learning research \u2013\u00a0from teaching software to play the strategy game Go better than any human\u00a0on the planet to creating artificial intelligence that mimicked the human brain’s ability to dream and even imagine future actions.<\/p>\n
Most of these breakthroughs, however, have seemed academic, with potential commercial uses seemingly years or even decades away. In the first year Alphabet owned the company, DeepMind\u00a0reported zero revenue.<\/p>\n
But when Google showcased its new digital assistant Wednesday\u00a0in its\u00a0flagship\u00a0Pixel\u00a0phone, laptop computer, and Google Home devices, the company\u00a0noted how much more natural its computer-generated voice had become.<\/p>\n
That more human-sounding speech, at least when the assistant speaks in English and Japanese,\u00a0is thanks to an algorithm DeepMind\u00a0invented last year and which it took from research\u00a0to full-scale commercial application inside twelve months.<\/p>\n
When DeepMind first published a 2016 paper on WaveNet, a new way of using software to generate speech,\u00a0tests showed that human listeners rated it more natural-sounding than existing technologies by a 50% margin.<\/p>\n
But the method was too computationally intensive. Even DeepMind\u2019s own researchers said the system was \u201cnot something we could deploy in the real world.\u201d<\/p>\n
Over the past year it found a way to make the\u00a0WaveNet algorithm 1,000 times faster while also allowing it produce even higher fidelity sound. This results in computer speech that human listeners judge to be almost indistinguishable from a real person\u2019s voice.<\/p>\n
These efficiency improvements \u2013 run using data centers equipped with Google\u2019s\u00a0new\u00a0computer chips\u00a0– have allowed the company to deploy WaveNet\u00a0into its new assistant.<\/p>\n
This clarity will be useful for DeepMind\u2019s bottom line. Since joining in 2015, Google\u2019s chief financial officer Ruth Porat has been on a mission to drive down costs across Alphabet, forcing some of Google\u2019s more left-field projects to show a path to profitability.<\/p>\n
DeepMind\u00a0continues to remain independent from its parent company, but its\u00a0contribution to Google\u2019s product launch is well timed. It reported its first-ever revenues \u2013 40 million pounds ($30.3 million) in 2016 \u2013\u00a0from products and services it supplied to other Alphabet companies, according to filings made public on the UK business registry Companies House on Monday.<\/p>\n
This is just one example of how DeepMind\u00a0is starting to help Google. Some others that DeepMind has been willing to talk about include supplying algorithms that have helped Google boost the energy efficiency of its data centers by 15%,\u00a0and also improvements to Google\u2019s core ad words product that DeepMind says it cannot detail.<\/p>\n
Of course, hiring all the brainpower to produce those algorithms means that DeepMind\u00a0is still shelling out far more cash than it takes in. The company reported that \u201cstaff and other related costs\u201d equaled 104 million pounds in 2016 and the\u00a0company reported an overall loss of\u00a0\u00a094 million pounds, nearly double the amount it reported the year before.<\/p>\n