Microsoft aims to advance its conversational AI push with Semantic Machines acquisition

[KRS]

While Microsoft has made a significant progress in personal assistant and artificial intelligence (AI) research over the last couple of years, the software giant recognizes the fact that it still has a long way to go in terms of maturing its conversational AI efforts. It's a field where intelligent assistants are able to understand human communication and carry on natural dialogue.

 

To teach machines to sound more like humans, Microsoft has acquired Semantic Machines, a conversational AI startup based in Berkeley, California. David Ku, corporate vice president and chief technology officer of Microsoft AI & Research, said in a blog post that Semantic Machines is focused on building conversational AI using machine learning for a more natural way of discovering and interacting with data and services.

 

Microsoft has been pioneering AI research initiatives, particularly in the fields of speech recognition and natural language understanding. In late 2016, the company launched the industry's first bot-as-a-service capability, called Azure Bot Service, meant to help developers integrate Microsoft Cognitive Services into their bots. Last August, Microsoft's conversational speech recognition technology also achieved a 5.1 percent word error rate, which is the lowest threshold set by IBM for human parity.

 

As part of the acquisition, Microsoft plans to create a conversational AI center of excellence in Berkeley in an effort "to push forward the boundaries of what is possible in language interfaces." The company will also incorporate Semantic Machines’ technology into its own AI advances and absorb the whole Semantic Machines team, including technology entrepreneur Dan Roth, UC Berkeley professor Dan Klein, Stanford University professor Percy Liang, and former Apple chief speech scientist Larry Gillick.

 

The acquisition also aims to advance Microsoft's digital assistant Cortana and social chatbot called XiaoIce, which the company claimed has achieved up to 30 billion conversations across several platforms in China, Japan, the United States, India, and Indonesia.

 

More details

 

< Here >