Figures and ancient books used to be like two distinct rivers. When they meet, what kind of efficiency and response can they produce? Following the requirements of "promoting the digitization of ancient books" and "actively carrying out the research and practice of ancient book text structure, knowledge systematization and intelligent utilization" put forward in the newly issued opinions on promoting the work of ancient books in the new era, the reporter interviewed domestic scholars engaged in the intelligent research of ancient books to explore what "butterfly changes" digitization can bring to the protection and research of ancient books. The silent classics moved A picture of thousands of miles of rivers and mountains flowing horizontally is marked with several ancient place names such as "Wushui", "Lianxi" and "Yushan". Hundreds of Confucian scholars with floating clothes are moving slowly on the picture, as if they are "on their way". Their journey from one place to another represents their learning process and their school. This is the exhibit submitted by Ma Yuan, a student majoring in artificial intelligence of Peking University, and her classmates to the first "Peking University Digital Humanities Exhibition" - H5 exercises completed in JavaScript, which is called "Song Yuan study plan inheritance visualization system". "A large number of ancient books like the Song Yuan study plan are too far away from our times. If you are not a professional research scholar, you may not remember to read it at all. We want to attract young people to understand ancient books through this form like a game interface." Ma Yuan said. At the exhibition site, what can also bring a sense of "jump" is the "Zhuxi chronology visualization system" produced by sang Yuchen and other students under the guidance of Shi Rui, deputy director of the research center of ancient Chinese history of Peking University. It uses GIS (Geographic Information System) technology to visually present the chronology of Zhu Xi in time and space. Readers can independently click and understand Zhu Xi's life of studying, traveling and making friends. Digital technology even gives ancient books the ability to study humanities, far more than making them move like games. "Digital Humanities represents the transformation of the research paradigm of Humanities and Social Sciences in the intelligent information environment, from traditional text driven to data-driven. The materials of humanities research, such as documents, catalogues and artifacts, can be transformed into some form of data, so that big data and artificial intelligence technology can also deal with them. Visualization is only a side effect brought by Digital Humanities, which makes it easy to understand academic achievements. And its deep logic , is the change of research paradigm. " Professor Wang Jun, director of the Digital Humanities Research Center of Peking University, told reporters. At the exhibition site, Wang linxu, a doctoral student under his guidance, showed the data mining results of "Song Yuan study case", "Ming Confucianism case" and "Qing Confucianism case"—— "Academic relationship network diagram", which uses regular expressions to make statistics on the relationship between the characters in the song and Yuan learning case and the Qing Confucianism case, has nine types of "disciples", "family studies", "private Shu", "homology", "scholars", "talking about friends", "making friends", "follower" and "others", and the frequency is clear at a glance. "Through the reconstruction of knowledge map, ancient books are no longer a mountain of characters, and the internal structure and semantic relationship in ancient texts can be clearly extracted and displayed in a short time." Wang Jun said. In addition to Chinese ancient books can be "counted", can foreign ancient books be "counted"? The answer is yes. In the digital research project of Italian poet Dante conducted by Cheng Mo, a teacher of the Department of Spanish and Portuguese languages at the school of foreign languages of Peking University, the rhythmic structures Valle (Valley), splale (shoulder) and calle (path) of the most frequently repeated three line poems in Divine Comedy have been accurately extracted. Not only "reading", but also "deduction" In the past, the study of classics mainly depended on masters. On the basis of reading a large number of documents, the master relies on his own memory and speculative ability to produce ideological research results, and then resort to the pen to convey them to the public in the form of words. The research of classics assisted by machine intelligence is based on data. With the intervention of machine intelligence, scholars can obtain the ability to process massive data instantly, Wang Jun analyzed. An article published by Liu Shi, a professor of the Chinese Department of Tsinghua University, and Yin Xiaolin, a full-time researcher of the Chinese Poetry Research Center of Capital Normal University, made a big data analysis of 100 classic ancient books from the pre Qin Dynasty to the Qing Dynasty, and found a lot. If we rely on manual statistics, such results are difficult to produce in a short period of time. One of the changes that digital brings to the study of classical books is the improvement of efficiency. "In the field of poetry research, senior scholars mainly analyzed and summarized the rhythm of Chinese classical poetry through examples. Later, there were manual annotation statistics and quantitative analysis statistics based on a large number of poetry. However, these research conclusions came from manual statistics, and the single research took a long time." Du Xiaoqin, a professor of Chinese at Peking University, recalled. Is there a software that can accurately mark the sound and rhythm format and rhythm degree of all Chinese classical poems with one click? Since 2004, Du Xiaoqin and others have begun to build a database of ancient Chinese phonology and ancient Chinese poetry text, recording the phonology of more than 10000 Chinese characters and poetry of more than 9 million words. On this basis, they developed the "sound and rhythm analysis system of Chinese classical poetry". This system can mark and statistically analyze the rhythm of Chinese classical poetry in large quantities. Using this system, Du Xiaoqin wrote many monographs and published many papers, such as the evolution from Qi Liang poetry to Tang poetry, the sound law of the Six Dynasties and the physique of Tang poetry. Having worked in the field of digitization of ancient books for many years, Wang Jun wants to do more than one-way knowledge extraction and information integration of ancient books. The automatic collation system of ancient books developed by Tang Xuemei, Yan Chengxi and other doctoral students under his guidance can automatically mark the sentence reading, person name, place name, official position, book title and time of ancient books through in-depth learning of algorithms and large-scale corpus training. The average accuracy of sentence reading is 94%, and the accuracy of named entity recognition in historical data is 98%. "Classical literature research supported by intelligent technology is one of the important directions of Ancient Book Research in the future." Wang Jun said. The "firewood" of the new atmosphere of humanities was born here "The stars last night, the wind last night, the spirit of thousands of years will be the same night. One laurel and smoke show, people are in the Qionglouyuyu." In a public speech, sun Maosong, a professor in the Department of computer science and technology of Tsinghua University, showed a poem to the audience. "Can you see that this is a collection of sentences extracted from four ancient poems? The key is, can you see that this is created by robots?" Asked sun Maosong. Through algorithm and deep learning, artificial intelligence has been comparable to human photography, painting, composition and poetry. Creativity, a unique field of human beings, is gradually being intervened by machines, which also raises some ethical questions - for example, can machines be called "art" through acquisition rather than the creation produced by human beings when they are full of emotion? The same problem is easy to arise in the field of Humanities and academic research after AI empowerment. Can the results produced after the machine intervenes in the research of various ancient books, such as various statistical data, visual "Atlas" or "page", be recognized as ideological humanistic research results? If so, how to quantify their academic value? "These should also be regarded as a form of achievements. In various academic circles, more and more attention is paid to data sets. Humanities such as history and literature based on ancient book research should not be despised, but should be paid more attention. Visualization itself can help scholars gain more insights on the one hand, and better spread to the public on the other. There are some ways that are difficult to achieve by traditional means, which is a new atmosphere of humanities The 'fuel fire' needs to be protected. " Yuan Xiaoru, a professor at the school of intelligence at Peking University, answered the reporter's questions. "Both the visualization results themselves and the communication effects of the results can be measured. Of course, although data-driven introduces intelligent technology into the humanities, the use and interpretation of data still need the intervention and guidance of humanities scholars." Wang Jun said. The newly issued opinions on promoting the work of ancient books in the new era requires that "strengthen the data circulation and collaborative management of ancient books, realize the aggregation and sharing of digital resources of ancient books", "support key units of digital ancient books to become stronger and better, and strengthen the management and open sharing of digital resources of ancient books". What are the reasons behind this? "Because the intellectualization of ancient books and the humanities and academic research based on them require a lot of capital investment. Computing tool platform, data resources, technical service team, etc. However, the financial strength of each research institution is inconsistent. The traditional research method that can produce a large number of results by relying on one or two scholars may not be applicable in the digital age. In order to make up for the academic difficulties caused by the difference of capital investment If there is a gap, it is necessary to strengthen sharing. " Wang Jun said. "Peking University can shoulder the task of building national infrastructure, share these facilities with the outside world, and help remote areas or places with insufficient academic resources to carry out research." Yuan Xiaoru said. The new movement of digital protection and utilization of ancient books has been played. (outlook new era)