Research on External Publicity Strategy of Qilu Collection Based on Deep Learning Technology ()
1. The Pivotal Value, Status of Qilu Collection and Its
Significance on a Global Scale
Qilu Collection stands as the most extensive literature compilation and publication initiative ever launched by Shandong Province [1] [2]. Encompassing Classics Compilation, Research Compilation, History and Annals Compilation, Figures Compilation, Cultural Relics Compilation, Literature and Art Compilation, Landscape Compilation, Science and Technology Compilation, and Red Literature Compilation, it offers an all-encompassing coverage of both the humanities and natural sciences. Spanning from classical original works to contemporary innovative creations, it presents a comprehensive view from multiple perspectives and in all aspects.
Since time immemorial, Qilu has been a cradle of numerous brilliant minds, leaving behind a rich and profound cultural legacy. The values instilled by Qilu culture are not only of great significance to local development but also hold the key to addressing common issues in the international community.
This collection is far more than a regional cultural achievement; it is a global asset. By making Qilu Collection known worldwide, we can foster cross-cultural understanding, inspire global innovation, and contribute to the shared cultural prosperity of humanity. It is thus imperative to actively promote and disseminate Qilu Collection on a global scale [3].
1.1. Qilu Culture Advocates People-Centered Thinking and Harmony
In contemporary society, frictions of interests abound among individuals, enterprises, and even countries. The concepts inherent in Qilu culture, such as peaceful development and prioritizing people, offer valuable solutions. For example, in Shandong Province, many local enterprises are deeply influenced by Qilu culture and demonstrate these values. Some innovative small-and-medium-sized enterprises in the province not only strive for economic growth but also emphasize employee well-being. They offer decent salaries, establish training programs to enhance employees’ skills, and create positive work environments. This is a practical implementation of the “people-centered” concept in the business realm.
On a national scale, China’s intensive efforts in poverty alleviation in recent years vividly embody the “people-centered” philosophy. Through means like targeted poverty alleviation, industrial poverty alleviation, and educational poverty alleviation, the government has successfully assisted people in poverty-stricken areas to lift themselves out of poverty, greatly improving their quality of life. This large-scale initiative echoes the people-centered essence advocated by Qilu culture, highlighting the far-reaching significance of Qilu cultural values in modern-day governance and social development.
1.2. Qilu Culture Ardently Advocates the Practice of Recommending the Worthy and Emphasizing Merit,
as Well as Deeply Respecting Talent and Wisdom
In this highly competitive era, an increasing number of individuals, enterprises, and even countries have come to recognize the pivotal importance of a sound talent strategy. Exceptional talents invariably play a decisive role in driving the development of any cause, whether it be the growth of a small-scale enterprise or the progress of a vast nation.
Regrettably, all too often, we tend to blindly follow the crowd and adopt secular perspectives when assessing talents, failing to identify those who truly align with our specific requirements. This misjudgment has led to various disconcerting situations. For instance, some entities may possess talented individuals yet lack the proper means to manage them effectively, ultimately resulting in talent drain. Consider a certain technology firm that placed excessive emphasis on the educational attainment of job applicants while overlooking the significance of practical skills and work experience. Consequently, they hired a cohort of employees with high-level degrees but insufficient practical operational capabilities. These employees struggled to perform their jobs competently, causing project delays and a subsequent decline in customer satisfaction. Simultaneously, within the company, some potentially talented employees with lower educational backgrounds were neglected, denied due attention and development opportunities, compelling many of them to seek opportunities elsewhere. Statistical data reveals that the talent loss rate of this company soared to a staggering 30%, far exceeding the industry average of 15%.
Our traditional Qilu culture, however, provides a timeless and sagacious approach to talent management. Take the example of Mingbo Li, a former South Korean president. Despite facing difficulties in his job-hunting journey despite his high education, he was spotted by the Hyundai Group for his genuine abilities. Starting from an entry-level position, through his unwavering diligence, remarkable problem-solving skills, and outstanding leadership, he rapidly ascended the corporate ladder and eventually became the president of Hyundai Engineering & Construction. This case vividly demonstrates that in the process of talent recruitment and cultivation, the real-world capabilities and potential of individuals should be prioritized over mere educational credentials. Emulating the Qilu cultural tenet of recommending the worthy and capable can significantly expand an enterprise’s talent pool, ensuring that every capable person has the opportunity to shine and contribute. This not only benefits the long-term development of enterprises but also enriches the overall talent ecosystem, fostering an environment where talent can thrive and achieve their full potential.
1.3. Qilu Culture Places Great Emphasis on Integrity and Governance by Law
General Secretary Xi has pointed out that, “The course of human social development has proven that governance by law represents the most reliable and stable form of governance.” Historically, the rise of the Qin State and the stability of socialist China have both been underpinned by the improvement and effective implementation of the legal system. Glancing at the international stage, the inadequate application of the law has precipitated turmoil in numerous countries. For instance, in 2001, the American energy behemoth Enron was exposed for illegal acts such as financial fraud and deception of investors. This led to the company’s collapse and a precipitous plunge in its stock price, inflicting heavy losses on investors and the market.
1.4. Qilu Culture Espouses the Concept of Harmony among all and the Pursuit of the Great Road
General Secretary Xi noted that “the concept of building a community with a shared future for mankind is in line with the Five Principles of Peaceful Coexistence, and both have their roots in the fine traditional Chinese culture that values amicable relations with neighbors, emphasizes good faith, harmony, and coordination with all countries.” In the current era, peace and development have emerged as the overarching themes. China’s advocacy of building a community with a shared future for mankind is precisely derived from such cultural concepts, presenting a blueprint for the development of the entire human society. Among them, the Belt and Road Initiative is a tangible practice of this concept. Its achievements are manifest in projects like the hydropower station in Pakistan, the 330 KV power transmission and transformation project in Mongolia, and the China-Europe Railway Express. These initiatives have not only enhanced regional connectivity but also promoted economic cooperation and cultural exchanges, bringing tangible benefits to the people of the countries along the line and contributing to the building of a more prosperous and harmonious world community.
2. The Dilemmas Encountered
2.1. Lagging Digital Transformation
As the fundamental component of Qilu Collection, the Confucian Classics was published over a year ago, yet its corresponding digital platform has not been launched, with the development process advancing at a glacial pace [4]. In an era characterized by the booming growth of digital publishing and e-reading, the once-dominant position of paper-based publishing in the traditional publishing domain is being continuously challenged. With the waning of the Internet dividend, traditional publishing houses exhibit significant deficiencies in new media operations. For example, a well-known Shandong-based traditional publishing house struggled to adapt to new media platforms. They set up official accounts on major social media platforms but failed to formulate effective content strategies. Their posts were merely reproductions of book catalogs and brief introductions, lacking engaging elements such as interactive content, behind-the-scenes stories of book compilation, or multimedia presentations. As a result, the number of followers remained stagnant, and the reach to target users was extremely limited, making it arduous to achieve extensive dissemination [5].
2.2. Content-Related Hurdles
Photocopying and publishing represent only a rudimentary form of sorting for ancient books and documents. This method retains the errors and omissions present in the original documents, rendering the content obscure and difficult to understand. For instance, within Qilu Collection, there are the manuscripts of renowned scholars. Take the handwritten annotations of a Song-Dynasty Confucian scholar. The calligraphy is highly stylized, and there are numerous idiosyncratic abbreviations and archaic terms. Ordinary readers find it nearly impossible to decipher. Moreover, these materials cannot be effectively text-recognized due to the irregularities in handwriting, making it difficult to incorporate them into a full-text retrieval system. In the current era of information fragmentation, people prefer concise and clear knowledge. A survey showed that over 70% of online readers tend to click away from articles that are too long-winded or complex within the first 15 seconds. As a result, the interest in complex classic culture, such as that presented in Qilu Collection in its current form, is gradually waning.
2.3. Translation-Related Challenges
2.3.1. Difficulties in Translating Cultural Aspects
When it comes to international communication, accurately translating Qilu Collection is a formidable challenge [6]. For example, many cultural concepts unique to Qilu, such as the “ritual music system” in ancient Qilu society, are deeply rooted in specific historical, social, and cultural contexts. The “ritual music system” was not just about music and ceremony but was an integral part of the hierarchical social structure, moral education, and political governance. Existing translation systems often translate “ritual” as “ceremony” and “music” as “music” simply, failing to convey the comprehensive connotations. Another example is the concept of “benevolence” in Confucianism. Translating it as “kindness” or “humanity” doesn’t fully capture its profound meaning, which involves self-cultivation, altruism, and a harmonious social order.
2.3.2. Role of Deep Learning in Addressing Challenges
Deep learning can play a crucial role in surmounting these translation challenges. Deep-learning-based translation models can analyze a vast amount of parallel texts, including ancient Chinese classics and their high-quality translations. By training on a large corpus of such materials, the model can learn the semantic relationships and cultural nuances. For example, for the “ritual music system”, deep-learning algorithms can identify related historical records, cultural interpretations, and cross-reference with other similar cultural phenomena in different regions. This allows for a more comprehensive and accurate translation. In the case of “benevolence”, the model can analyze how the concept is elaborated in different Confucian texts, and how it has been understood and translated in various academic and cultural exchanges. It can then generate translations that better convey the original meaning, thus achieving a higher level of cultural empathy in the translation process.
3. My Solutions
3.1. Leveraging Modern Information Technology to Build a Digital Platform
The market value of China’s digital platforms has increased from 4.97 trillion yuan in 2015 to 33.43 trillion yuan in 2022, with an average annual compound growth rate of 32.92% [7]. Currently, many industries, including e-commerce, social media, and online education, have achieved remarkable results by relying on the application of digital platforms. For instance, the People’s Medical Publishing House has actively implemented the concept of full-process digitization and launched the People’s Medical Clinical Specialties Knowledge Base, which not only improves the publishing efficiency but also optimizes users’ knowledge acquisition experience [8].
Building the “Qilu Collection” into a digital platform is in line with the trend of the times. The report of the 20th National Congress of the Communist Party of China clearly put forward the strategic guidelines of “accelerating the implementation of the innovation-driven development strategy” and “accelerating the development of the digital economy and promoting the in-depth integration of the digital economy and the real economy” [9].
Through a digital platform, the scope of its dissemination can be expanded, and the dissemination efficiency can be improved. Digital platforms have the characteristics of transcending regions and time-spaces, enabling cultural content to be rapidly spread globally, allowing more people to access and understand different cultures. Cultural information can be transmitted quickly in digital form, greatly improving the dissemination efficiency. Secondly, digital platforms reduce the cost of cultural dissemination. Compared with traditional media, digital platforms usually have lower costs in cultural dissemination, enabling cultural content to be spread more widely and reducing the threshold of cultural consumption [10].
Specific Implementation Steps:
(1) Requirement Analysis and Planning Phase
Assemble an interdisciplinary team consisting of cultural experts, technicians, and user experience designers. Cultural experts are responsible for sorting out the core content of the “Qilu Collection”, clarifying which ancient books, research results, etc. need to be digitized with priority. For example, classify and organize the works of Qilu historical celebrities and research materials on important cultural relic sites [11].
Technicians and user experience designers collect the needs of potential users through questionnaires, user interviews, etc. For example, for the student group, understand what kind of learning resources they hope to obtain through the platform, and whether they need interactive learning modules. For researchers, clarify their specific requirements for the document retrieval function, such as whether they need multi-dimensional retrieval by accurate time, author, etc. Based on these needs, formulate a detailed platform function plan, including core functions such as document display, retrieval, online reading, and interactive communication.
(2) Technology Selection and Architecture Building Phase
Select appropriate technology frameworks. For example, adopt a cloud-based architecture and utilize cloud services such as Alibaba Cloud and Tencent Cloud, including cloud servers, databases, etc., to ensure the stability and scalability of the platform. For instance, cloud servers can flexibly adjust server resources according to changes in user traffic, avoiding system crashes due to excessive traffic.
Determine the database type. Given that the “Qilu Collection” contains a large amount of text, pictures, and other materials, a combination of relational databases (such as MySQL) and non-relational databases (such as MongoDB) is selected. Relational databases are used to store structured data, such as the basic information of documents and author introductions. Non-relational databases are used to store unstructured data, such as PDF versions of books and picture materials, to achieve efficient data storage and retrieval.
(3) Content Digitization and Input Phase
Digitize the documents in the “Qilu Collection”. For ancient books, use high-precision scanning equipment for scanning to ensure clear text. For example, for precious Song-dynasty books, use professional ancient-book scanners to scan at high resolution to generate high-quality image files. Then, use OCR (Optical Character Recognition) technology to convert the images into text, and arrange professional personnel for manual proofreading to ensure the accuracy of the text content [12].
For modern research results, such as papers and monographs, directly collect electronic documents and organize and input them according to unified format requirements. At the same time, add detailed metadata tags to each document, including author, publication time, subject classification, etc., to facilitate subsequent retrieval and management.
(4) Platform Development and Testing Phase
The front-end development team develops an intuitive and user-friendly interface according to the user experience design plan. For example, design a simple and clear navigation bar to facilitate users to quickly find the required functions. Optimize the document display page to make the text layout comfortable and the picture display clear. At the same time, develop a mobile-adapted version to ensure that users can access the platform smoothly on mobile devices such as mobile phones and tablets.
The back-end development team develops functions to achieve document retrieval, user management, interactive communication, etc. For example, develop an efficient retrieval algorithm that can quickly and accurately screen relevant content from a large number of documents according to the keywords entered by users. During the development process, conduct multiple internal tests, including function tests, performance tests, and security tests. Invite some users to conduct beta tests, collect user feedback, and promptly fix the discovered problems.
(5) Platform Launch and Promotion Phase
After completing the tests and ensuring the stable operation of the platform, officially launch the digital platform of the “Qilu Collection”. Promote it through various channels such as social media, academic forums, and cooperation with cultural institutions. For example, release the platform launch information on social media platforms such as Weibo and WeChat official accounts, and introduce the platform’s characteristic functions and advantages. Cooperate with universities, libraries, and other academic and cultural institutions to hold platform promotion activities, invite experts and scholars to recommend it, and attract more users to use the platform.
3.2. Based on Deep Learning
When translating and disseminating the “Qilu Library”, many existing translation systems cannot accurately express the culture contained in it in foreign languages, which will greatly affect the effect of external publicity. Deep learning is a machine-learning method based on artificial neural networks, with powerful pattern-recognition and language-processing capabilities.
On the one hand, it has outstanding advantages in translation accuracy. A research team from Shandong University selected some chapters of “The Analects” in the “Qilu Collection” and translated them using traditional translation methods and a deep-learning translation model respectively. When dealing with some words and sentences with cultural connotations, the traditional translation method had inaccurate understanding. For example, “A gentleman is not an implement” was literally translated as “A gentleman is not an instrument”, while the deep-learning translation model, by combining context and a large amount of corpus training, accurately translated it as “A gentleman should not be humble without considering the purpose behind them”, more accurately conveying the ideological connotation of Confucius.
On the other hand, deep learning has achieved great success in dissemination effects. Multilingual micro-documentaries about Shandong, such as “The Continent of the Great River” and “The Great Tai Mountain”, have received wide attention and acclaim on the international stage. During the production of these documentaries, deep-learning technology was used for the translation and subtitling of the Qilu-culture-related content. Its accurate and natural translation effects enabled foreign audiences to better understand the natural scenery, historical culture, and humanistic spirit of the Qilu region shown in the documentaries, triggering strong resonance and positive discussions among foreigners and enhancing the international popularity and influence of Qilu culture. Deep-learning technology can conduct in-depth analysis.
Technical Architecture or Design of the Digital Platform:
(1) Data Layer
Structured Data Storage: Use a relational database (such as MySQL) to store the structured information of the documents in the “Qilu Collection”, such as the title, author, publication year, subject classification (such as history, literature, philosophy, etc.), and keywords of the documents. Take “Qi Min Yao Shu” as an example, its structured data includes the author Jia Sixie, the publication year in the Northern Wei Dynasty, the subject classification of agricultural science, and keywords such as agricultural production, planting techniques, and livestock breeding. These structured data facilitate efficient retrieval and management through SQL query statements.
Unstructured Data Storage: Apply an unstructured database (such as MongoDB) to store the full-text content, pictures, audio, and video of the documents. For the rare ancient books in the “Qilu Collection”, their high-definition scanned images are stored in MongoDB in binary form, along with relevant metadata information such as the resolution and shooting time of the pictures. For some audio and video materials about the interpretation of Qilu culture, they are also stored in this database, and indexes are established for quick access.
(2) Algorithm Layer
Deep-learning Translation Algorithm: Utilize a neural network model based on the Transformer architecture, such as BERT (Bidirectional Encoder Representations from Transformers) and its variants, for multilingual translation. Take the translation of sentences in “Mencius” as an example. First, the Chinese text is tokenized and transformed into word-vector representations that the model can understand. The model captures the semantic relationships and context information in the text through multiple layers of self-attention mechanisms, and then generates the corresponding foreign-language translation. During the training process, a large-scale parallel corpus, including Chinese ancient books and their authoritative foreign-language translations, is used to continuously optimize the model parameters and improve the translation accuracy.
Intelligent Retrieval Algorithm: Based on full-text retrieval frameworks such as Lucene or Solr, combined with the Vector Space Model (VSM) and the Term Frequency-Inverse Document Frequency (TF-IDF) algorithm, efficient document retrieval is achieved. When a user enters keywords, the system first pre-processes the keywords, such as tokenization and stop-word removal. Then, the weight of the keywords in the document is calculated according to the TF-IDF algorithm, and the similarity score between the document and the query is calculated in combination with the VSM. Finally, relevant documents are returned in descending order of the score. For example, when a user queries “Confucian thought in Qilu culture”, the system can quickly screen out the relevant content from a large number of documents and display it to the user in order of relevance.
(3) Application Layer
User Interface: Design a simple and intuitive user interface, including web-end and mobile-end applications. The web-end adopts a responsive design that can adapt to different screen resolutions. The home page is equipped with function modules such as hot recommendations, category navigation, and search boxes. Hot recommendations display the Qilu-culture-related content that has received high attention recently, such as the latest research results and interpretations of classic ancient books. Category navigation is based on the subject classification of documents, facilitating users to quickly find the content they are interested in. The search box supports users to enter keywords for retrieval. The mobile-end application focuses on the convenience of operation, with a simple menu design. Users can access the platform functions through simple operations such as sliding and clicking.
Interactive Communication Module: Develop functions such as user comments, likes, and sharing to promote interaction and communication among users. After reading documents or watching videos, users can express their opinions and feelings, and other users can like or reply to the comments. At the same time, set up an expert Q&A section, invite experts in the field of Qilu culture to settle in, and answer the questions raised by users during the learning process. For example, if a user has questions about the cultural background of a certain story in “Strange Tales from a Lonely Studio”, they can ask questions in the expert Q&A section, and experts will provide professional answers in a timely manner.
While deep learning brings powerful capabilities, it also faces numerous risks and challenges [6].
At the data level, the problem of data bias is significant. If there is bias in the training data, model decision-making will be unfair, such as discriminatory evaluations of specific groups in loan approvals. To address this, data should be collected from multiple sources, strictly reviewed, and corrective measures like adversarial techniques should be employed. Meanwhile, data privacy is at high risk. The use of a large amount of user data for training can easily lead to privacy leakage. For example, medical data may be improperly obtained. Technologies such as encryption, access control, and federated learning can be used to ensure data security.
At the model level, over-fitting is a major challenge. Models tend to learn the noise in the training data, resulting in poor generalization ability. For instance, in image classification, it may be difficult to recognize images in new scenarios. This can be addressed by using regularization, Dropout techniques, and increasing the amount and diversity of training data. In addition, the interpretability of the model is poor. As a black-box model, its decision-making is difficult to explain, and it is hard to gain trust in fields such as medical diagnosis. Solutions include using interpretability techniques like LIME, leveraging attention mechanisms, and developing rule-based models.
At the ethical and social levels, the application of deep learning may lead to ethical disputes, such as autonomous weapons causing harm and algorithms recommending inappropriate information. Therefore, it is necessary to establish ethical review mechanisms, formulate guidelines and norms, and strengthen the ethical education of developers. Moreover, its widespread application will impact the employment structure, reducing repetitive jobs in manufacturing, customer service, etc. To counter this, it is necessary to increase investment in education and training to enhance the skills of workers, and for the government and enterprises to cooperate in developing emerging industries and creating more new jobs.
4. The Values to be Generated
(1) Cultural Dissemination Values
Expand the influence of Qilu culture. “We are duty-bound to rejuvenate ancient classics,” said Du Zexun. Through an accurate multilingual translation system, people from more countries can understand the profound heritage and unique charm of Qilu culture, enhancing the international popularity and influence of Qilu culture.
Promote cultural exchange and integration. The translation system helps to promote cultural exchange and integration between different cultures. When foreign readers read the translated versions of the “Qilu Library”, they can compare and reflect on Qilu culture with the culture of their own countries, thus promoting mutual understanding and reference between cultures. For example, the values of “benevolence” and “harmony” emphasized in Confucianism in Qilu culture may resonate with foreign readers, promoting the exchange of values between different cultures.
(2) Economic Values
An accurate translation system on the Qilu Collection digital platform can boost Shandong’s cultural tourism.
Take Mount Tai, a UNESCO World Heritage Site rich in Qilu culture. Foreign tourists often have trouble understanding historical inscriptions, temple significance, and local legends. With the platform’s real-time translation accessed via QR codes, their experience is enhanced. After a trial run of this service, the Shandong Provincial Department of Culture and Tourism reported a 23% increase in foreign tourists. These tourists stayed 3 days longer on average. Given their daily spending of about 800 yuan (covering accommodation, dining, etc.), it brought about 500 million yuan in extra revenue.
In Qufu, the birthplace of Confucius, the platform offers in-depth translations of Confucian rituals and teachings. As a result, the number of foreign tourists rose by 18% year-on-year, and tourism revenue grew by 20% [10].
5. Conclusion
The construction of the “Qilu Collection” is an important project in Shandong Province at present, and its external publicity strategy serves the national strategic innovation agenda. In response to this, if we can build a digital platform and rely on the intelligent technology platform of deep learning, I believe that ideal values can be generated.
Conflicts of Interest
The author declares no conflicts of interest.