Taiwan's Microsoft 23rd first held Microsoft AI Day, Microsoft Dean Hon said, machine intelligence is to multiply, this is the strength of the machine, depth learning from a broad sense is a pattern recognition, machine stronger than pattern recognition, such as 10 phone numbers, chess and so on, are similar to the principle. In contrast, hon analysis, human intelligence is to simplify the complex, find a simplified law, according to the law to solve the problem, we have been discussing ' intelligence ', ' intelligence ' the highest realm is to simplify. The son said: ' The ginseng! My way is consistent. ' This consistent, is to simplify, as long as a truth can break all the work, this is the human intelligence, rather than the machine to multiply system. Hon pointed out that AI development is at a turning point, the computer is also constantly sophisticated, through more information learning, in the near future can deal with more complex and routine tasks; However, to really solve the problem requires super smart Si (Super Intelligence), is to combine AI and hi human intelligence (Human Intelligence) can be achieved. Hon stressed, from Einstein's time and space environment background to observe, Einstein nearly from zero data put forward gravitational waves, 100 years later, human only to measure the gravitational wave, the current so-called AI is to large data, no data, no AI, at least at this stage, There is no AI can be based on small data or 0 data to make predictions, the future may have, but in a short period of time it is difficult to make a great application, as for the application of large data will be very quickly. Hon believes that creativity cannot be taught and AI has a long way to go in the same capacity as science fiction or film, because our understanding of human intelligence is still very limited, not to mention the AI to human intelligence. Hon further explained that we were the first generation of humans to live with AI, we believe that human intelligence will evolve with AI as well, so there is no need to worry too much about the development of AI, human beings can use the technology of the past to face AI, by AI Auxiliary to enhance capacity and create higher business value. AI development is facing four challenges, including the industry's demand for AI talent surge but the problem of serious supply imbalance; The entrepreneur has a cognitive gap between the prospects of AI and how to use AI; The overall environment of AI platforms and tools is not enough to support enterprises and individuals easy and smooth application; and related AI data is not easy to obtain and difficult to extract valuable data from. Microsoft puts forward the five levels of AI intelligence hierarchy, and continues to study its applications to help enterprises break through these difficulties. The five capabilities of AI intelligence hierarchy structure from computational and memory, perception and visual auditory ability, cognitive insight and inference planning to decision-making ability, creativity and sense of relativity, ultimately have a considerable intelligence ability. In the most basic computing and memory architecture, Microsoft to the AI popularization to the purpose, in the Azure platform to build AI framework, AI operations, data platform, robot framework, cognitive services, Advanced machine learning. In the second layer of perceptual and auditory ability architecture, ability to learn in depth and have image recognition, such as the five major AI cognitive service APIs that Microsoft offers enterprise applications, including visual, speech, language, knowledge, search, such as Microsoft How-old.net can identify people's visual age, ' Twins Or not "The face similarity test, Microsoft launched a photo identification Ai ' Captionbot ', as long as the upload of any image, can automatically identify and explain the content of this photo, including the photo of the characters who, this is a kind of posture, landscape location or status." In the framework of the third level of cognitive insight and inference planning to decision-making capabilities, Microsoft wants to accelerate the data to insights by building simple and powerful insights and communication experiences. such as the Microsoft AI Robot small ice, with host programs, creative poetry, singing, writing news and other capabilities. In the fourth level of creative architecture, from the production of poetry or lyrics and fragments of melody, to be able to differentiate and can create, such as the Microsoft Small Ice Poetry ' sunshine lost Windows '. Image style transformation such as the Microsoft PIX Camera App service, you can change the photo as an art masterpiece. Image style conversion can also have the same function and effect. The semantic level of style transfer, can identify the image is original or synthetic map, can accurately convert the desired style, but also can be converted to the original art works very similar style. Microsoft's AI team, including business and product research specialists, scientists, engineers, design and ethics experts, economics and Data science experts, Microsoft for the AI platform, in the technologically advanced and powerful cloud, provide an elastic platform to enhance developers and enterprises to connect with Microsoft's AI services and tools, to play a rapid innovation capabilities . Microsoft integrates AI in its own product services, such as the Windows ten-Hello biometric, Cortana voice assistant and Information Security Center for automated response and decision-making Office 365 (translate), say (read aloud), read (automatically identify and depict the picture content), and write ( Dictation voice to text); Azure's perception service, robot framework, natural language Understanding services Luis and image identification and description; Bing Intelligent Search and Skype ... and other products. In addition, 2017 Microsoft AI Day invites major industry representatives to share with people the ways in which AI actually applies to industry, such as vice president of NVIDIA Global operations and Taiwan District general manager Chillimen, to share observations and recommendations on the use of AI by Taiwanese companies to upgrade their industries; Dr. Churen, chief executive, integrated Microsoft Cloud CRM system dynamic 365, through the intelligent robot dialogue to analyze customer identity, further recommend the most suitable products to customers, improve the efficiency of precision marketing.
An intelligent CEO Tang Guolen, sharing the steps of the monitoring industry to achieve AI, from the browser plus a large number of equipment management and analysis of the cloud, to the end of the cloud face recognition popularization, as well as Microsoft Azure Cloud to help build a public cloud in private clouds solutions, to achieve AI in the technology Fangzhong, long-term care The last mile of the entertainment industry; Rui Jie International Co., Ltd. Dr. Wang Zhizhong, sharing How to apply AI industry health diagnosis, through the process, quality, equipment and other state of effective detection and integration, to assist in the practice of Industrial 4.0 service production and maintenance decisions.
