WaveNet Machine-Generated Speech from DeepMind Eclipses Competitor Technology

July 13, 2017

The article on Bloomberg titled Google’s DeepMind Achieves Speech-Generation Breakthrough touts a 50% improvement over current technology for machine speech. DeepMind developed an AI called WaveNet that focuses on mimicking human speech by learning the sound waves of human voices. In testing, the machine-generated speech beat existing technology, but is still not meeting the level of actual human speech.

The article expands,

Speech is becoming an increasingly important way humans interact with everything from mobile phones to cars. Amazon.com Inc., Apple Inc., Microsoft Inc. and Alphabet Inc.’s Google have all invested in personal digital assistants that primarily interact with users through speech. Mark Bennett, the international director of Google Play, which sells Android apps, told an Android developer conference in London last week that 20 percent of mobile searches using Google are made by voice, not written text.

It is difficult to quantify the ROI for the $533M that Google spent to acquire DeepMind in 2014, since most of their advancements are not extremely commercial. Google did credit DeepMind with the technology that helped slash power needs by 40%. But this breakthrough involves far too much computational power to lend itself to commercial applications. However, Google must love that with the world watching, DeepMind continues to outperform competitors in AI advancement.

Chelsea Kerwin, July 13, 2017


Got something to say?

  • Archives

  • Recent Posts

  • Meta