Today, the MIDC·2020 Xiaomi Developer Conference was held. Xiaomi released a number of AIoT-related products including the XiaoAi 5.0 voice assistant and the mobile deep learning framework MACE Micro. After the meeting, Xiaomi Group General Manager of Artificial Intelligence Department, Ye Hangjun and other executives were interviewed by NetEase Technology and shared Xiaomi smartphone + AIoT latest development strategy.
At this conference, Xiaomi upgraded Xiao Ai’s previous positioning to an intelligent life assistant. For this, Ye Hangjun said that Xiaomi announced a new strategy in August this year – smartphone + AIoT will build a smarter life. From a strategic point of view, this is the company’s new strategy.
Wang Gang mentioned that the original intention of Xiao Ai was to lower the threshold for users to use the product. Due to the wide variety and quantity of smart devices nowadays, using smart devices may become a burden for users. Xiao Ai, as a software capability on smart hardware devices, can make the use of smart devices simpler. In this upgrade, the core of all-scenario intelligent collaboration is that no matter how complex the type of smart device is, users can simplify all the complex logic behind the device by simply using the voice interaction of Xiao Ai.
There is a need to continuously break through industry difficulties
Wang Gang introduced that the current R&D and breakthrough directions for Xiao Ai are still focusing on its basic functions. It includes wake-up, speech recognition, and understanding of users’ natural semantics, etc. And it will continue to deepen and refine several functions. For example, is the voice that directly interacts with users more natural and emotional? In addition, a series of industry difficulties such as false awakening need to be constantly broken.
In addition, Wang Gang said that Xiao Ai 5.0 includes several major functions. They are full-scene collaboration, active dialogue reminders, emotional sounds, multi-modal fusion interaction, and intelligent life assistants. They will decentralize most of the functions to all the devices built-in Xiao Ai. And because some of these functions require a certain amount of time and energy to adapt, starting from the end of this year, they will be decentralized for different devices.
Smartphone + AIoT is the biggest landing scene of artificial intelligence
Ye Hangjun introduced that Xiaomi’s AI department is currently divided into two parts. One part is the research and development of AI products represented by Xiao Ai. And the other part is the AI laboratory.
Ye Hangjun said that from the level of cooperation with business departments, because Xiao Ai itself has a user-visible product form, it will be relatively clear, including several major devices such as smartphones, speakers, and TVs, followed by AIoT devices. Among them, the AI department provides the SDK, and different equipment teams will discuss product planning with the AI team. Then, they will optimize the AI function through technology and product evaluation, as well as the special production needs of the business department. The AI laboratory is mainly responsible for the R&D and upgrading of functions such as vision, speech, acoustics, natural language understanding, and machine learning.
In addition, Ye Hangjun said that smartphone + AIoT is the largest scenario for artificial intelligence landing. And the long-term accumulation of scenario data will form the core advantage of enterprise AI technology.