三年多来，"识典古籍"平台累计上线4.7万部古籍资源，月服务用户超240万人，日均检索量达35万人 次，总访问量突破1.47亿次。该中心主任王军介绍，平台不仅有阅读功能，还有古籍整理功能，引入人 工智能技术，可实现文字识别、自动标点、专名提取、智能校勘等。 截至目前，来自全国近1500所高校的2万名学生，以及1.8万名社会志愿者共3.8万人参与了平台古籍智能 整理。北大等高校已将"识典古籍"平台引入课程教学，让学生可在实践中培养古籍整理技能。 转自：北京日报客户端 北京大学启动建设三年多的"识典古籍"已成为全球规模最大的古籍智能数字化整理和阅读平台，《永乐 大典》等4.7万部古籍资源实现免费查阅。被誉为"中国中古时代的百科全书"的敦煌写卷已启动"收 录"。 敦煌写卷即1900年在莫高窟藏经洞发现的文献，目前分藏于全球数十家收藏机构。此次启动整理的敦煌 写卷原件藏于法国国家图书馆，数字版于2025年11月拷贝赠予北大数字人文研究中心。这些重要文献将 陆续亮相"识典古籍"平台。 来源：北京日报 ...

论坛围绕数智时代古籍研究的转型与创新展开讨论，古委会副秘书长吴国武介绍了高校系统在古籍整理 研究方面的数智化实践，他认为高校正在逐步打破"依赖专家学者人工整理"的传统模式，越来越多地将 智能工具深度融入学科建设与人才培养。北京大学中文系教授杨海峥则结合古典文献学发展脉络，深入 分析了数智技术驱动下学科转型的趋势与核心方向，为古籍研究的现代化升级提供学术参考。 长期以来，字节跳动公益持续发挥技术能力，助力古籍的保护与传承。与北京大学联合打造的"识典古 籍"，在字节跳动豆包大模型支持下，不仅高效助力古籍全流程整理，还提供便捷智能的查阅与研究功 能，每月服务用户超240万人，平台总访问量突破1.47亿。字节跳动公益相关负责人表示，团队将持续 提升深度研究与知识挖掘的智能化水平，推动更多古籍实现资源共享，欢迎广大图书馆、文献馆免费接 入"识典古籍"，以满足相关需求。 1月10日，由全国高等院校古籍整理研究委员会(以下简称古委会)主办，北大中文系古典文献教研室、 北大数字人文研究中心与字节跳动公益联合承办的"数智时代古籍数字化前沿论坛暨'我用AI校古 籍'(2025年)总结会"在北京举行。教育部社科司、北京大学、抖音集 ...

Core Insights - The launch of the "Deep Research Assistant" feature on the ancient book digitalization platform "Shidian Ancient Books" represents a significant advancement in AI application for in-depth research of ancient texts [1][2] - This feature utilizes advanced AI technology to automate the research process, enhancing efficiency and depth in the study of ancient literature [1][2] Group 1: Feature Overview - The "Deep Research Assistant" is an autonomous AI tool capable of conducting comprehensive research, from planning to report generation, significantly improving the efficiency of ancient book studies [1] - The platform currently includes 36,000 ancient texts, and the assistant can analyze these texts to extract valuable insights [1] Group 2: User Feedback and Development - The development of the "Deep Research Assistant" was driven by feedback from scholars who found traditional research methods cumbersome and time-consuming [2] - The assistant allows users to input research topics, which it then uses to perform exhaustive analysis and generate reports, including visual data representations [2] Group 3: Reliability and Trustworthiness - All conclusions drawn by the "Deep Research Assistant" are based solely on the content of the ancient texts within the "Shidian Ancient Books" platform, ensuring reliability [2] - The assistant employs a unique "citation chain" mechanism to enhance credibility, assigning exclusive IDs to each segment of text analyzed [2] Group 4: Future Development and Applications - The "Deep Research Assistant" is currently in a trial phase, with ongoing user and expert feedback being used for continuous improvement [2] - This tool also serves as a "repository of inspiration" for content creators, facilitating the generation of accurate and engaging historical materials [2]

Core Insights - The launch of the "Deep Research Assistant" feature on the "Shidian Ancient Books" digital platform marks a significant advancement in the field of ancient book research, utilizing AI technology for in-depth analysis and research support [1][3] Group 1: Product Features - The "Deep Research Assistant" is an autonomous AI tool that can conduct comprehensive research, from planning to report generation, significantly enhancing the depth and efficiency of ancient book studies [1][3] - It is designed to go beyond traditional Q&A models by creating research plans, extracting information from digital ancient texts, and generating structured reports with charts [3][6] - The assistant's conclusions are based solely on the ancient texts within the platform, ensuring reliability and credibility through a unique "citation chain" mechanism [6] Group 2: User Experience and Feedback - The development of the "Deep Research Assistant" was influenced by feedback from scholars who found traditional research methods cumbersome and time-consuming [3][8] - In testing scenarios, the assistant demonstrated high accuracy, closely matching the number of examples found in scholarly papers, and received positive feedback for its clarity and innovative perspectives [6][8] - The feature is currently available for trial, with ongoing user and expert feedback being used for continuous improvement and optimization [8] Group 3: Market Position and Potential - This feature represents the first deep knowledge mining attempt in the digital ancient book sector, filling a gap in AI applications for in-depth ancient research [3] - The assistant serves as a resource for content creators and enthusiasts, providing quick access to accurate historical materials and making ancient studies more engaging and efficient [8]

Core Insights - The launch of the "Deep Research Assistant" feature on the ancient book digitalization platform "Shidian Ancient Books" represents a significant advancement in AI application for in-depth research of ancient texts, enhancing research efficiency and depth [2][3] Group 1: Product Features - The "Deep Research Assistant" utilizes advanced AI technologies, including ByteDance's Doubao large model, to conduct comprehensive analysis and research on the platform's collection of 36,000 ancient texts [2] - Unlike traditional Q&A bots, this assistant can autonomously create research plans, extract knowledge, organize data into tables, and generate research reports complete with charts [2] - The feature addresses the challenges faced by scholars in manually extracting information from vast amounts of literature, streamlining the research process significantly [3] Group 2: User Experience and Feedback - The development of the "Deep Research Assistant" was driven by feedback from scholars using the platform, indicating a need for more efficient research tools [3] - The assistant can serve as a "repository of inspiration" for content creators, quickly generating accurate and engaging historical materials, thus supporting various forms of ancient text revitalization [3] - It also acts as an online mentor for enthusiasts, making the study of ancient texts more engaging and efficient while maintaining a professional standard [3]

Core Insights - The 2024 ByteDance Corporate Social Responsibility Report highlights the company's efforts in using technology to enhance societal goodwill, particularly through the digitization of ancient texts [1] Group 1: Ancient Texts Digitization - ByteDance has been actively promoting the digitization of ancient texts using AI technologies, which is crucial for preserving Chinese cultural heritage [1][4] - The "Chinese Ancient Texts Protection Plan" estimates that there are approximately 200,000 types of ancient texts and 500,000 versions, with many important texts already digitized [1] - The "Shidian Ancient Texts" platform, launched in October 2022, has made over 16,000 ancient texts available for free, with a monthly user base exceeding 2.4 million [2] Group 2: Community Engagement and Crowdsourcing - The "I am a Text Editor" initiative has mobilized over 10,000 students from more than 1,000 universities to participate in the digitization process, resulting in the proofreading of over 600 million characters [3] - The "I Use AI to Edit Ancient Texts" campaign aims to further engage the public and students in the digitization efforts [3] Group 3: Technological Advancements - The platform utilizes advanced AI technologies such as OCR, automatic punctuation, and structure organization, significantly improving efficiency in text processing by at least five times [4] - ByteDance has committed 25 million yuan to support the digitization of Confucian texts, which will also be made available on the "Shidian Ancient Texts" platform [4] Group 4: Cultural Revitalization - ByteDance is leveraging new technologies to revitalize traditional culture, with significant engagement in non-material cultural heritage and museum collaborations [5] - The company has seen a 356% increase in orders for non-material cultural heritage products, indicating a growing public interest [5] Group 5: Future Prospects - The digitization of ancient texts and the creative transformation of traditional culture are seen as essential for cultural preservation and development in the digital age [5][6]