The research on artificial intelligence has made great progress in recent years, but it is still at the beginning of the period that all computers want to understand human communication. Most of today's robots and intelligent voice assistants can respond to simple commands, such as broadcasting weather forecasts, playing songs, but cannot understand the meaning and dialogue. To make communication more effective, an assistant needs to be able to engage in natural conversations rather than just responding to commands. This is called "conversational AI."
On May 20, Microsoft announced the acquisition of Semantic Machines, Inc. of Berkeley, Calif., which has developed a revolutionary new method to establish dialogic AI. They are the ability to use machine learning to enable users to discover, access information, get services and interact with them in a more natural way and with greater efficiency.
Member of Semantic Machines Company
Lei Fengnet also learned that Semantic Machines is composed of many pioneering leaders in the field of conversational artificial intelligence, including Dan Roth, a science and technology entrepreneur, as well as globally outstanding and innovative natural language AI researchers, Dan Klein, a professor at the University of California, Berkeley, and Stanford. University professor Percy Liang and former Apple Siri chief speech scientist Larry Gillick.
David Ku, CTO of the Microsoft Artificial Intelligence and Microsoft Research Division, said that for more than two decades, Microsoft has been working on research and breakthroughs in building basic modules of conversational AI, such as speech recognition and natural language understanding. Our goal is to extend the "vision" of the computers around us to the world that they can see, to hear and understand people's conversations. In 2016, we took a big step toward the goal of achieving session computing, introduced a framework for robot development, and released pre-built cognitive services for artificial intelligence assistants, including speech recognition and natural Language understanding. Today, more than 1 million developers use Microsoft's Microsoft Cognitive Services and over 300,000 developers use the Azure Bot Service, all of which help make computing more interactive.
Microsoft is also further enhancing the conversational AI capabilities of digital assistant Cortana and social chat robots like Xiao Bing. Currently, Xiao Bing has conducted more than 30 billion conversations with an average of 30 minutes per session. These conversations come from 200 million users in China, Japan, the United States, India, and Indonesia. Through Xiao Bing and Cortana, Microsoft has made breakthroughs in the field of speech recognition, and it has recently become the first company to add full-duplex voice sensing to conversational AI systems. The dialogue between people and the system is more convenient and natural.
David Ku also said that as Microsoft acquires Semantic Machines, Microsoft will establish a dialogue AI center in Berkeley to promote the development of natural language technology. Combining Semantic Machines' technology with Microsoft's artificial intelligence technology provides a more powerful, more natural and efficient user experience, and also brings session computing to a new level.
Since conversational AI is an innovation in human-computer interaction and an entry point for services in the AI era, so with the fierce competition of smart speakers, one of the carriers of conversational AI, Lei Fengnet learned that in addition to Microsoft, Amazon, Google, Apple and China Baidu attaches great importance to dialogue AI. Apple Siri was one of the first voice conversation assistants, but its product experience has been unsatisfactory. In Amazonia, in 2012, Amazon engineers began investing in the development of conversational speech assistants, and gradually evolved into Alexa. Google launched Google Assistant, a voice intelligence assistant, in May 2016, and in September of the same year it also acquired the API.AI team that provides developers with Siri-style conversational robot development tools. Baidu introduced DuerOS. DuerOS is a dialogue-type artificial intelligence system developed by Baidu Dupei Division. Baidu hopes to use voice as an entry point to create key nodes for future smart homes and everything. The giants are actively investing. Who will lead the conversational AI?
Lei Feng network compiler, viaMicrosoft Blog