In 2018, Wechat has more than a billion monthly active users. In China, almost everyone with a mobile phone is using Wechat. Wechat has become an indispensable part of life, and the AI technology used in Wechat will also become a great help in people's lives. With the arrival of the wave of artificial intelligence, the major Internet companies have "gone to sea" one after another. As China's leading Internet value-added service provider, how does Tencent lay out this new format?
I. organizational structure
1. Tencent AI Lab: focus on the basic research and application exploration of artificial intelligence, and provide AI technical support for Tencent's product business
Founded in 2016, Tencent AI Lab is committed to developing general artificial intelligence research, exploring the connection between the virtual world and the real world, and developing AI+ medicine. AI Lab is an important part of Tencent's layout in the field of artificial intelligence. As an enterprise laboratory, it has comprehensive AI basic research capabilities, focusing on basic research including computer vision, speech recognition, natural language processing and machine learning. According to Tencent's business, it puts forward four application directions of game, content, social and tool platform AI.
two。 Robot Lab Robotics X: Tencent AI Industrial basic support Department, exploring the connection between the virtual world and the real world
On March 15, Yao Xing, vice president of Tencent, announced at the second academic Forum of Tencent AI Lab that Tencent will set up a robot laboratory "Robotics X", which, together with AI Lab, will become the double basic support department of Tencent's AI industry. Yao Xing said, "I believe that with the establishment and future growth of Robotics X, we will certainly be able to develop a market similar to the Internet industry in the early PC era and the mobile Internet industry in the smartphone era. "
3. AI teams of major business groups: responsible for AI application research
Social Network Group (SNG) YouTu Laboratory is Tencent's top machine learning research and development team, focusing on image processing, pattern recognition, and deep learning; Wechat team and Hong Kong University of Science and Technology jointly established the "Wechat-Hong Kong University of Science and Technology artificial Intelligence Joint Laboratory", WHAT Lab research perspectives include robot, NLP learning, data mining, speech recognition. Wechat Beijing R & D Center's research interests include speech recognition, image recognition, speech synthesis, audio fingerprinting, semantic understanding, data mining, related to Wechat voice to text, voiceprint recognition, shake and other business; Tencent Intelligent Computing and search Lab (TICS LAB) was established in 2015, focusing on four major research directions: search technology, natural language processing, data mining and artificial intelligence.
II. Product application
1. Tencent News: establish user interest graph for intelligent personalized recommendation
In Tencent news apps (such as Kuaibao), Tencent Video, QQ Music, National singing and app stores (such as YingYongBao), AI can help these products better understand users' interest graphs, thus making more intelligent and personalized recommendations, so that users can experience interesting content more efficiently. Digital content services on Tencent's various platforms will benefit from this smarter recommendation technology.
two。 Pitu: practical Application of face recognition Technology
The detection rate of face recognition technology in Tencent and YouTu Lab ranks first according to the world-class FDDB standard. The "put on military uniform" H5 campaign in moments precisely adopts the face fusion technology pioneered by Tencent and Pitu in the industry. At the same time, Tencent's face recognition technology can also be used for user identity online authentication, used in finance, government affairs, public welfare and other fields.
3. Tencent Miying (MIAIS): application in early screening and diagnosis of diseases
Using the deep learning technology provided by Tencent AI Lab, early screening and diagnosis of diseases are carried out in medical images, mainly supporting lung cancer screening, esophageal, cervical and breast cancer screening, as well as observation of diabetic retinopathy. In addition, Tencent Miying uses NLP natural speech processing technology to analyze medical institutions and information, and through the case information, the pathological information is structured, and the diagnostic risk and treatment effect are predicted by the model.
4. Tencent's unique skill: Chinese version of AlphaGo
"unique Art" was independently developed by a 13-member team of Tencent AI Lab for nearly a year. Learning mainly includes human chess score database and machine self-playing. Its algorithm is based on the two cores of strategy network and value network, and innovatively greatly improves the accuracy of value network. In the future, Tencent will continue to increase investment in AI, open up the "unique skill" technology, and help the technical research of go AI, so as to better inherit go culture.
5. Tencent Wenzhi: one stop to meet the needs of NLP, transcoding, extraction and data crawling
Tencent Wen Zhi Natural language processing (Natural Language Processing) is based on parallel computing, distributed crawler system, combined with unique semantic analysis technology, one station to meet the needs of NLP, transcoding, extraction, data crawling and so on. Based on Wen Zhi API, it can realize the functions of search, recommendation, public opinion and mining, and support customized semantic analysis scheme at the same time.
III. Investment layout
Data source: knowledge chain data
In 2017, Tencent conducted investment mergers and acquisitions covering 12 related areas, including animation, games, short videos and movies, reading and content payment, social networking and tools, education, health care, life services and e-commerce, finance, automobile transportation, corporate services and artificial intelligence. Tencent's investment in artificial intelligence is mainly focused on automobile transportation, not only investing in NIO Inc. Automobile with Baidu, Inc., but also investing in Tesla, Inc., the benchmark of smart car, sharing car-hailing Ola Cabs and Go-Jek, and flying car Lilium Aviation.
Among them, Kukai is a section under Skyworth that specializes in the production of smart TVs and their operating systems. Tencent acquired a 7% stake with a heavy sum of 300 million yuan in June 2017, with a valuation as high as 4.2 billion yuan. Being optimistic about the "living room economy" of the home Internet is a major reason for Internet companies to invest in the color TV industry. In addition, as a kind of sharing economy, small electric charging treasure of small Power Technology is placed in a fixed portable battery cabinet in a restaurant, entertainment and leisure place and other small scenes in the form of a scene. Consumers can consume and eat in the store while using small electricity, scan the QR code to pay for the corresponding length of mobile phone charging.
IV. Competitors
1. Baidu, Inc.: more than 2000 artificial intelligence patents
As the earliest artificial intelligence company among the three BAT companies, Baidu, Inc. hopes to recover the backward situation and realize "corner overtaking". By focusing on the application scene, it promotes the commercialization of AI technology. In addition to using AI to drive the growth of old business, it also uses AI in the financial field, deploying DuerOS, Apollo, focusing on automobile, home and other scenes to form a new long-term business.
two。 BABA: the "ET City brain" project is settled in Suzhou and Hangzhou.
BABA has launched "industrial AI" programs such as ET brain in urban, industrial, retail, financial, automobile, family and other scenes. Take ET City brain as an example. In Hangzhou, it can not only increase the speed of vehicles on congested roads by 15.3%, but also give priority to special vehicles such as 120 ambulances in Xiaoshan area, saving half the time to arrive at the scene. ET City brain, as one of the first members of the "artificial Intelligence National team", compared with Baidu, Inc. 's Apollo and Tencent medical images, is the only solution that has landed and achieved results.
3. Alphabet Inc-CL C: strategic layout of Alphabet
Alphabet Inc-CL C turned the artificial intelligence extracted from search into Alphabet Inc-CL C's brain and applied it to various fields. For example, AlphaGo in the field of go, Google car in the field of automobile, Google house in the field of medicine, and Google genetics; in the level of genetic testing, so on Alphabet, Alphabet Inc-CL C's ambition is to turn the success of one Alphabet Inc-CL C into 26.
(editor / Zhang Huiyu)