Video: Microsoft democratizes AI adding it to core products and services.
Microsoft has acquired Semantic Machines, an AI startup from Berkeley, California, that will now help the Redmond company accelerate its natural-language AI efforts.
The move follows Google’s Duplex demo that wowed developers by phoning a hair salon to book an appointment, all apparently without the human on the receiving end of the call ever knowing she was talking to a machine.
Semantic Machines’ team includes a number of former executives of Nuance Communications and engineers who worked on Siri and Google Now.
Microsoft says it will combine Semantic Machines with its own AI research to “take conversational computing to a new level”, enabling chatbots to go beyond simple commands and queries to carrying on a full conversation using what Microsoft calls “full-duplex voice sense”.
Microsoft in April claimed to have made a breakthrough on this front with its popular XiaoIce chatbot in China, which has had 30 billion conversations, averaging up to 30 minutes each, with 200 million users across platforms in China, Japan, the US, India, and Indonesia.
Noted talent the acquisition gives Microsoft include Apple’s former chief speech scientist of Siri, Larry Gillick, UC Berkeley professor Dan Klein, and Stanford University professor Percy Liang, who helped create the language-understanding AI behind Google Assistant.
Microsoft didn’t disclose the value of the acquisition. Semantic Machines was founded in 2014 and secured backing of $21m from Bain Capital Ventures and General Catalyst.
The company was working on machine-learning technology to enable bots to understand the meaning and context of a conversation, even as it moved in different directions.
The acquisition will give a boost to Microsoft’s Cognitive Services and Azure Bot Service that collectively are used by 1.3 million developers, according to Microsoft.