Microsoft buys Semantic Machines to make AI sound more human
Cortana and Xiaolce to see benefits of acquisitionBy Greg Synek
Microsoft has purchased startup company Semantic Machines in an effort to make artificial intelligence bots sound more human. The Berkeley, California-based business focuses on contextual understanding of conversation.
Previously, the firm has worked with Apple on speech recognition technology for Siri. Semanitc Machines is lead by professor Dan Klein of UC Berkeley and professor Percy Liang of Standford University in addition to Apple's former chief speech scientist Larry Gillick.
Microsoft has been working on speech recognition and natural language processing for nearly two decades now. As Cortana has gained a more prominent role in recent years, Redmond is aiming to improve the accuracy and fluency of its assistant.
In addition to Cortana, Xiaolce is Microsoft's Chinese chatbot that has just recently achieved full duplex capabilities. The bot is able to listen to other callers, provide near real-time responses, and speak over the caller if an interjection is needed. Xiaolce has held more than 30 billion conversations in total with an average conversation lasting around 30 minutes.
By acquiring Semantic Machines, Microsoft will be setting up a dedicated center for AI research in Berkeley to take on the challenges of speech. Conversational computing will be expanded upon and further integrated into future Microsoft products.
Even though Microsoft is trying to up its game in language processing and conversation handling, there will still be fierce competition with Amazon and Google for AI supremacy. Apple's Siri assistant is quickly becoming a dated tool that lacks the more modern capabilities achieved by all other major voice systems.