On the morning of March 31, the 102nd issue of the Publishing Forum (Special Lecture 5 of epidemic situation and Publishing Series) hosted by the school of information management and Publishing Research Institute of Nanjing University was broadcasted online at Tencent conference and station B platform. The forum invited the deputy general editor of China earth press Co., Ltd. and Geological Publishing Co., Ltd., the executive general manager of China earth press and China earth digital media Zhang Xinxin, general manager of Beijing Science and technology culture Co., Ltd. and chairman of financial think tank, gave a lecture entitled "publishing transformation in the era of artificial intelligence". Professor Yang Haiping, vice president and doctoral supervisor of Nanjing University Publishing Research Institute, presided over this lecture. More than 7200 people, including leaders, colleagues, teachers and students from the national press and publication administration, industry, University and Research Institute, watched the lecture online, and the response was very warm, which was well received by everyone.
At the beginning of the lecture, president Zhang recommended Jerry Kaplan's "the age of artificial intelligence". He pointed out that our publishing industry is now in the trend of transformation and upgrading, and the opportunities are fleeting, so we should grasp them in time and transform gracefully. In his book "Ji guangpianyu: Publishing transformation in the age of artificial intelligence", president Zhang pointed out that it is difficult to complete the transformation and upgrading by relying on the department system, and only the company system can be competent for the great mission of publishing transformation. Mr. Zhang told us in detail how to "gracefully transform" publishing in the era of AI from four aspects: Ai Ai Ai and publishing transformation, AI concept and development process, intelligent publishing process, news publishing + AI.
No.1
AI artificial intelligence and publishing transformation
First of all, president Zhang explained the main contents of the new generation of artificial intelligence development plan issued by the State Council in July 2017. The plan made it clear that artificial intelligence entered a new stage and positioned it as a new focus of international competition, a new engine of economic development and a new opportunity for social construction. There are eight key generic technologies, six of which are closely related to the press and publishing industry, including knowledge computing engine and knowledge service technology, cross media analysis and reasoning technology, group intelligence key technology, virtual reality intelligent modeling technology, autonomous unmanned system intelligent technology, natural language processing technology.
The related press and publishing formats include ar publishing, VR publishing, knowledge service, composite publishing process reengineering, news and publishing big data, intelligent robot application, etc. With the development of digital publishing entering the deep water area and entering the stage of high-quality development, the digital transformation and upgrading of future publishing will focus on four areas: knowledge service, online education, full copyright operation and intelligent publishing.
Knowledge computing engine and knowledge service technology. We should focus on breaking through the core technologies of knowledge processing, deep search and visual interaction, realize the automatic acquisition of continuous increment of knowledge, have the ability of concept recognition, entity discovery, attribute prediction, knowledge evolution modeling and relationship mining, and finally form a multi-source, multi-disciplinary and multi data type cross media knowledge map covering billions of entity scale.
Virtual reality intelligent modeling technology. We should focus on breaking through the intelligent behavior modeling technology of virtual objects, improving the sociality, diversity and interactive fidelity of intelligent object behavior in virtual reality, and realizing the organic combination and efficient interaction between virtual reality, augmented reality and artificial intelligence.
Natural language processing technology. We should focus on breaking through the core technologies of natural language grammar logic, character concept representation and deep semantic analysis, promote the effective communication and free interaction between human and machine, and realize the intelligent understanding and automatic generation of natural language with multi styles, multi languages and multi fields.
No.2
The concept and development of artificial intelligence
The concept of artificial intelligence has been popular for several years. In 2017, it became one of the top ten buzzwords in Chinese media, and is recognized as one of the three cutting-edge technologies (genetic engineering, nanoscience, artificial intelligence) in the 21st century. So what is artificial intelligence? Different scholars have given different definitions. President Zhang thinks that the most appropriate definition of AI is "a computer program that makes reasonable actions according to the perception of the environment and obtains the maximum benefit". In short, AI is to give machines intelligence.
The development of artificial intelligence has experienced three upsurges.
The first AI wave occurred from 1956 to 1976. The Dartmouth conference held in 1956 hoped to establish artificial intelligence as a scientific task and complete path, marking the formal birth of artificial intelligence.
The second AI wave occurred from 1976 to 2006. The popular research fields of the second wave include: speech recognition, speech translation, neural network, expert system, etc. The most famous technology is "expert system", which is usually composed of six parts: human-computer interface, knowledge base, reasoning machine, interpreter, comprehensive database and knowledge acquisition.
The development of expert system has gone through three stages, and it is transiting to the fourth generation.
The first generation of expert system is characterized by high specialization and strong ability to solve special problems. The second generation of expert system: it is a single discipline professional and application-oriented system. The third generation of expert system: it is a multi-disciplinary comprehensive system, which adopts a variety of artificial intelligence languages, a variety of knowledge representation methods, a variety of reasoning mechanisms and control strategies. The fourth generation expert system: on the basis of summarizing the design method and implementation technology of the first three generations of expert system, it has started to use the latest artificial intelligence technology, such as large-scale multi expert cooperation system, multi-disciplinary collaborative problem solving and parallel reasoning, knowledge acquisition and learning mechanism of artificial neural network.
General manager Zhang pointed out that at present, publishing has built a knowledge base in various professional fields, which is an expert system in the second stage. The next step is to realize the construction of interdisciplinary knowledge base.
The third AI wave started in 2006 and continues to this day. The hot research fields of the third wave include big data, deep learning, etc. Deep learning based on big data, training big data, and summarizing the knowledge or rules that can be used in similar data by computer. Big data and deep learning are two "pillars" that push the third wave of artificial intelligence to a climax.
President Zhang pointed out that the path of the second wave and the third wave is not the same. The second wave tried to use deductive methods to exhaust knowledge in a field, and the third wave tried to use inductive methods to continuously learn knowledge.
It is worth mentioning that the three waves all have marked events in the field of sports competition. The first wave in the field of sports is a landmark event: in 1962, IBM's Arthur Samuel developed a checkers program to beat a blind Checkers Master. The landmark event of the second wave in the field of sports is that on May 11, 1997, the parallel computer "dark blue" developed by IBM in the United States defeated Kasparov, who has been the world chess king for 12 years. The landmark event of the third wave in the field of sports is: in March 2016, alphago defeated Li Shishi, the world go champion and professional nine stage player, with the principle of in-depth learning, with a total score of 4:1. At the 2017 Internet Summit in Wuzhen, China, alphago beat the world's top go player Ke Jiejiu 3-0.
No.3
Intelligent publishing process
After combing the concept and development process of artificial intelligence, president Zhang then specifically introduced how to examine the publishing industry from the perspective of artificial intelligence, which would bring changes to the publishing process. At present, editing, review, proofreading, printing, storage and transportation, sales and other links still rely on a large number of human resources investment, and replacing human resources investment with a large number of automated and intelligent systems will become the main trend. In the era of artificial intelligence, publishing process reengineering, which takes group intelligence as the concept guide and knowledge system construction as the core, will shine brilliantly.
Plan
The biggest inspiration of AI for planning is group intelligence, which improves topic selection planning by means of collaborative innovation. Group intelligence has been applied to various fields in real life, such as Wikipedia based on group editing, open-source software developed by group, knowledge sharing of public Q & A, innovation of public intelligence, sharing economy of crowdsourcing, etc.
Mr. Zhang cited the poetry collection "sunshine lost the glass window" created and published by Microsoft's artificial intelligence "Xiaobing" and the ancient poetry created by artificial intelligence "Weiwei" from the speech and language experimental center of Tsinghua University. Without considering the quality of the works, the emergence of "Xiaobing" has brought challenges to the traditional topic selection planning and writing of publishing enterprises. "Weiwei" can even pass the "Turing test", which means that "Weiwei" can no longer recognize whether the author is an artificial intelligence or a natural person.
Writing
In terms of intelligent writing, "writer" artificial intelligence technology platform wordsmith, in 2013, the number of news articles automatically written by the machine reached 300 million, exceeding the output of all major news organizations; in 2014, more than 1 billion news articles have been written.
The advantages of machine writing lie in the support of big data corpus, more objective reporting, faster publishing speed, and lower error volume. At the same time, it also promotes the transformation of journalists to the direction of in-depth news reporting. The disadvantages lie in the insufficient depth of the news, lack of highlights and focus, and limited to information broadcast, sports finance and other fields.
Proofreading
In terms of proofreading, news publishing enterprises have built a series of intelligent proofreading systems, such as automatic error correction system, sensitive word recognition and screening system and system compilation system, based on the theories and technologies of big data intelligence, group intelligence and natural language processing. The emergence of a series of intelligent proofreading systems, such as intelligent proofreading system, sensitive word recognition system, AI audit system and rich media AI audit system, has greatly increased the number of intelligent proofreading systems Improved process efficiency. However, at present, the reserve of professional knowledge of each discipline of the intelligent proofreading system is not enough. The real future lies in the inclusion of the domain vocabulary of each industry.
printing
In the aspect of printing, the main development direction of intelligent printing is the greenization of new printing consumables, the digitalization before printing, the networking and high efficiency of equipment, the automation of printing process, the networking of technical services and the informatization of management system.
issue
When it comes to the transformation of distribution, the most popular one is the online Red live broadcast sales. According to incomplete statistics, vivia, the first sister of the live broadcast, can sell 50000 copies in two hours. The new retail of online live broadcast is an important aspect of the publishing and distribution process.
On the other hand, the performance of AI participating in the distribution is that the intelligent distribution can build an automatic accurate user portrait evaluation system based on the customer relationship system, and realize the accurate matching of printing scale and distribution quantity based on this, forming a data return loop, so as to achieve the goals of optimizing supply, reducing return rate, statistical analysis and accurate delivery 。
No.4
Press and publication + AI: intelligent product service scenario
In addition to the artificial intelligence reshaping the publishing process, president Zhang also shared with you seven intelligent product service scenarios that AI can provide for the press and publishing industry.
News publishing big data
In the era of intelligence, the deep transformation of the press and publication industry will inevitably be accompanied by the upgrading and evolution from digitalization and fragmentation to digitalization and intelligence. As an important part of the construction of news publishing big data, the construction of professional publishing big data has achieved initial results in China: audit, public