中文    English

Journal of library and information science in agriculture

   

Development and Construction of Metadata Specifications for AI Models

JIANG Enbo1,2, QIN Yu1,2   

  1. 1. Chengdu Library and Information Center, Chinese Academy of Sciences, Chengdu 610041
    2. School of Economics and Management, University of Chinese Academy of Sciences, Beijing 100190
  • Received:2025-06-22 Online:2025-10-29

RichHTML

0

PDF (PC)

0

Abstract:

[Purpose/Significance As artificial intelligence (AI) systems are being widely deployed across diverse domains such as education, healthcare, and public governance, the absence of standardized metadata specifications has led to fragmented descriptions, inconsistent documentation, and difficulties in model evaluation and reuse. This study aims to address the pressing issues of opacity, lack of interpretability, and poor traceability in current AI models, which have increasingly become obstacles to the development of transparent and responsible AI. To overcome these challenges, this study proposes the establishment of a unified metadata specification for AI models to enhance their discoverability, transparency, interoperability, and reusability, thereby advancing the development of trustworthy AI and facilitating effective model governance. [Method/Process Grounded in metadata quality assessment theory and lifecycle theory, the study adopted a combination of research methods, including literature review, comparative analysis of existing specifications, and questionnaire surveys.We first conducted a systematic examination of domestic and international practices related to AI model metadata specifications to identify representative standards, frameworks, and implementation approaches. Through comparative analysis, the study investigated the structure, element organization, and semantic relationships of different specifications, highlighting their similarities, differences, and areas for improvement. Meanwhile, a targeted questionnaire survey was administered to researchers, developers, and practitioners to explore user awareness, perceptions, practical experiences, and specific needs regarding metadata specification and interoperability. Based on these findings, the study ultimately proposed a lifecycle-oriented framework for metadata specification construction, ensuring that it aligns with the key stages of AI model development, deployment, evaluation, and governance. [Results/Conclusions The findings reveal that, although users generally recognize the importance of metadata specifications for AI models, they are unaware of of the existing specifications. The current AI model metadata specifications have significant shortcomings in terms of element naming, structural organization, and descriptive granularity. These shortcomings hinder the effective sharing and reuse of model information. In response, the study proposed a comprehensive metadata framework encompassing key entities such as models, datasets, algorithms, technical features, performance evaluations, risks and ethics, legal information, and related resources, as well as the semantic relationships among these entities. The research concluded that establishing a unified metadata specification for AI models not only contributes to effective information management and cross-platform interoperability, but also serves as a critical infrastructure that links technology, ethics, and governance. As the metadata specification system matures and gains wider industry adoption, AI models will become increasingly controllable and trustworthy. This will promote a more regulated, collaborative, sustainable and integrated AI ecosystem.

Key words: AI models, model transparency, metadata specification, AI

CLC Number: 

  • G251

Table 1

Publicly released metadata information of current mainstream large language models"

维度 GPT-4[9] BERT[10] Claude[11] LLaMA[12]

文心一言

（ERNIE Bot）[13]
所属公司 OpenAI Google Anthropic Meta 百度
发布时间 2023.3 2018 2023 2023.2 2023.3
模型架构 Transformer Transformer Transformer Transformer Transformer
参数规模 未公开 BERT-base：1.1亿 未公开

LLaMA-1: 7B/13B/33B/65B

LLaMA-2: 70B

 未公开
训练数据 仅声明互联网大规模文本、书籍等[14] 英文维基百科、BookCorpus数据[15] 仅声明使用互联网公开数据和第三方企业许可的数据[16] The Pile、BooksCorpus、Common Crawl、C4（Colossal Clean Crawled Corpus）、arXiv等[17] 仅声明覆盖中文互联网、维基百科、百度知识图谱、图片、语音等数据[18,19]
训练方法 大规模预训练+RLHF等对齐技术，未公开具体的技术细节[13] 自监督预训练-下游任务微调[15]

自监督预训练

强调Claude的训练方法注重对齐，采用类似RLHF的人类反馈训练[20,21]

 自监督预训练[17] 自监督预训练-中文场景多任务微调[18]
训练成本 未公开 未公开 未公开 未公开 未公开
版本和更新历史 GPT-1（2018）、GPT-2（2019）、GPT-3（2020）、GPT-3.5（2022）、GPT-4（2023） BERT-base、BERT-large同期发布

Claude 1（2023年初）

Claude 2（2023年中）

LLaMA-1（7B~65B，2023.2）

LLaMA-2（70B，2023.7）

 文心一言（2023年3月首发）、随后迭代多个版本
道德与隐私声明 “使用政策”“道德守则”以及隐私政策[7,22] 应遵循通用的Google AI道德准则[23] 强调用户数据保护与安全，对敏感请求设限[24,25] 未发布明确的道德与隐私声明 发布个人信息保护规则、智能体功能服务条款、用户协议等[26-28]

Table 2

Development of metadata specifications for AI models"

时间 组织 规范 适用模型类型 主要贡献与特点 涉及的实体关系
2018 Google Model Card 机器学习模型 提出了标准化的模型文档格式，记录机器学习模型的关键信息，推动模型透明度 训练关系、评估关系、产生关系
2018 W3C ML Schema 机器学习模型 提出用于描述机器学习模型的规范，促进不同平台与系统之间的互操作性 训练关系、评估关系
2019 Hugging Face Model Card 深度学习模型、大语言模型 专注于NLP领域，强调模型的复用性和社区贡献，进一步推动模型透明度 训练关系、评估关系、衍生关系、开发关系、反馈关系、产生关系
2019 IBM AI Factsheet 机器学习模型、深度学习模型 记录AI系统的性能、偏见、可靠性、安全性等关键指标，提高AI系统的透明度与可审计性 衍生关系
2020 Amazon AI Service Cards 机器学习模型、深度学习模型 为云服务中的AI模型提供详细元数据描述，帮助用户更好地理解和使用AI服务 衍生关系、产生关系
2021 OpenAI System Card 大语言模型 提供模型的详细信息，促进用户对模型的理解和信任 更新关系、训练关系、评估关系、产生关系

Table 3

Model card description elements"

数据项 数据项（中文） 描述内容
Model Details 模型详细信息

名称和版本、类型

开发团队和发布日期

框架与算法、参数、公平性约束等信息

获取更多信息的论文或其他资源

引文详细信息

许可证

向何处发送有关模型的问题或评论

Intended Use 预期用例

预期用途

预期用户

范围外的用例

Factors 因素

相关因素

评价因素

Metrics 指标

模型性能测量

决策阈值

变化方法

Evaluation Data 评估数据

数据集

动机

预处理

Training Data 训练数据

数据来源及统计特征

数据采样策略与分布

Quantitative Analyses 定量分析

根据评估指标提供评估模型的单一结果

根据评估指标提供评估模型的交叉结果

Ethical Considerations 道德考虑 模型开发中的道德考虑因素，向利益相关者提出道德挑战和解决方案
Caveats and Recommendations 注意事项与建议

潜在误用风险

用户操作建议

Fig.1

Model card: Toxicity detector"

Table 4

Metadata description information for AI models"

元数据项

Model Card

（Google）[29]

ML Schema

（W3C）[30]

AI Factsheet

（IBM）[31]

AI Service Cards

（Amazon）[32]

System Card

（OpenAI）[33]

Model Card

（Hugging Face）[34]

 覆盖情况
模型基本信息 模型名称 √ √ √ √ √ √ 6
模型版本 √ √ √ 3
开发团队 √ √ √ √ 4
发布时间 √ √ 2
模型架构 √ √ √ 3
模型算法 √ √ 2
模型参数 √ √ √ 3
任务类型 √ √ √ √ 4
应用场景 √ √ √ 3
训练数据 数据集名称 √ √ √ √ √ 5
数据来源 √ √ 2
数据特征 √ √ √ √ 4
处理策略 √ √ √ 3
评估数据 数据集名称 √ √ √ √ 4
数据特征 √ √ 2
处理策略 √ 1
评估指标 √ √ √ √ √ √ 6
技术特征 输入特征 √ √ 2
输出特征 √ √ 2
训练技术 √ 1
计算设施 √ √ 2
伦理与风险考虑 伦理与风险审查 √ √ 2
数据隐私 √ 1
偏见与公平性 √ √ √ 3
安全性 √ √ 2
反馈途径 √ √ √ 3
覆盖情况 50% 53.8% 57.7% 30.8% 42.3% 53.8%

Table 5

Examples of metadata field variations"

元数据字段

Model Card

（Google）

ML Schema

（W3C）

AI Factsheet

（IBM）

AI Service Cards

（Amazon）

System Card

（OpenAI）

Model Card

（Hugging Face）
模型基本信息 Model Details Model Model Information Overview Introduction Model Details
应用场景 Intended Use Intended Domain Intended use cases and limitations Uses
训练数据 Training Data Dataset Training Data Model Data & Training Training Details
评估指标 Metrics Model Evaluation Performance Metrics Performance Expectations Observed Safety Challenges and Evaluations Evaluation
偏见与公平性 Bias Fairness and Bias Bias, Risks, and Limitations
反馈途径 Contact Information Further Information Model Card Contact

Table 6

Distribution of respondents' personal characteristics"

个人特征 占比/%
年龄 18~25岁 19.51
26~35岁 65.85
36~45岁 9.76
46岁以上 4.88
职业 学生 19.51
教师 19.51
科研人员 26.83
企业员工 34.15
机构性质 高等院校 31.71
研究机构 29.27
公司企业 36.59
其他性质 2.44
学历 本科及以下 12.20
硕士研究生 56.10
博士研究生 31.71
行业 金融 2.44
医疗 9.76
教育 17.07
电商 17.07
信息与通信技术 46.34
其他行业 7.32

Fig.2

Statistics on channels for learning about AI models"

Fig.3

Statistics on stages of AI model usage"

Fig.4

Statistics on factors influencing decision-making"

Fig.5

Radar chart of respondents' familiarity with six specifications"

Fig.6

Statistics on respondents' perception of barriers to using AI models"

Fig.7

Respondents' perception of shortcomings in existing metadata specifications for AI models"

Fig.8

Respondents' needs for AI model description specifications/tools"

Fig.9

Entities and relationships in AI model descriptions"

Table 7

Model entity classification"

实体类别 描述内容 支撑字段举例
模型 标识信息、开发时间、模型类型、模型架构、模型参数等信息 Model Description、Model type
数据 模型涉及的数据集的来源、类型、格式及其数据处理、数据标注等信息 Data、Training Data、Testing Data
算法 算法名称与类型、算法原理、优化策略、超参数配置等信息 Algorithm
作者 模型的开发团队与数据集的创建者 Authors
基础设施 模型需要的开发环境、训练环境与部署环境等软硬件配置 Compute Infrastructure
技术特征 模型的技术属性、训练配置与优化技术等信息 Technical Specifications
性能 模型的各种评估/验证过程与结果、性能指标等 Factors、Metrics、Performance Expectations
应用场景 模型可适用的领域、场景以及实际应用案例 Uses、Task、Intended Use Cases and Limitation
风险 模型可能存在的数据隐私、社会伦理、模型偏见等法律与伦理风险，以及风险缓解措施等 Bias, Risks, and Limitations
法律信息 知识产权声明、许可协议、隐私信息、法律合规性与其他可能与法律相关的信息等 License、Licensing Information、Privacy、Intellectual Property
相关资源 其他与模型相关的论文及论文引用、文档与教程等链接 More Information、Model Sources、Quick Link
联系点 用于用户与开发者之间的反馈与交流，可服务于模型的更新与迭代，包括客户支持邮箱、社区论坛链接、问题反馈链接等联系性信息 Contact
[1]
БУСЛОВА M K, 王续琨. 模型化方法与模型的认识功能[J]. 世界科学, 1982(5): 47-51.
БУСЛОВА M K, WANG X K. Modeling methods and the cognitive functions of models[J]. World science, 1982(5): 47-51.
[2]
GAMBELLA C, GHADDAR B, NAOUM-SAWAYA J. Optimization problems for machine learning: A survey[J/OL]. arXiv: 1901.05331, 2019.
[3]
HU Z T, XING E P. Toward a "standard model" of machine learning[J/OL]. Harvard data science review, 2022, 4(4): 1-15.
[4]
HU X, CHU L Y, PEI J, et al. Model complexity of deep learning: A survey[J/OL]. arXiv: 2103.05127, 2021.
[5]
ZHAO H Y, YANG F, SHEN B, et al. Towards uncovering how large language model works: An explainability perspective[J/OL]. arXiv: 2402.10688, 2024.
[6]
ZHAO H Y, CHEN H J, YANG F, et al. Explainability for large language models: A survey[J/OL]. arXiv: 2309.01029, 2023.
[7]
DANG Y K, HUANG K C, HUO J H, et al. Explainable and interpretable multimodal large language models: A comprehensive survey[J/OL]. arXiv: 2412.02104, 2024.
[8]
QIN J, YU B. Metadata in trustworthy AI: From data quality to ML modeling[C]//International Conference on Dublin Core and Metadata Applications. Daegu: DCMI, 2024.
[9]
OpenAI. GPT-4 is OpenAI's most advanced system, producing safer and more useful responses[EB/OL]. [2025-03-02].
[10]
Google. BERT[EB/OL]. [2025-03-02].
[11]
Anthropic. Claude[EB/OL]. [2025-03-02].
[12]
Meta. Llama[EB/OL]. [2025-03-02].
[13]
百度. 文心一言(ERNIE Bot)[EB/OL]. [2025-03-02].
[14]
OPENAI, ACHIAM J, ADLER S, et al. GPT-4 technical report[J/OL]. arXiv: 2303.08774, 2023.
[15]
DEVLIN J, CHANG M W, LEE K, et al. BERT: Pre-training of deep bidirectional transformers for language understanding[J/OL]. arXiv: 1810.04805, 2018.
[16]
你们如何在模型训练中使用个人数据？[EB/OL]. [2025-01-26].
[17]
TOUVRON H, LAVRIL T, IZACARD G, et al. LLaMA: Open and efficient foundation language models[J/OL]. arXiv: 2302.13971, 2023.
[18]
ZHANG Z Y, HAN X, LIU Z Y, et al. ERNIE: Enhanced language representation with informative entities[J/OL]. arXiv: 1905.07129, 2019.
[19]
百度AI模型“文心一言”新鲜体验[EB/OL]. [2025-01-26].
[20]
OpenAI. GPT-4[EB/OL]. [2025-01-26].
[21]
BAI Y T, KADAVATH S, KUNDU S, et al. Constitutional AI: Harmlessness from AI feedback[J/OL]. arXiv: 2212.08073, 2022.
[22]
OpenAI. Privacy policy[EB/OL]. [2025-01-26].
[23]
Google. Our principles[EB/OL]. [2025-01-26].
[24]
Claude. Privacy policy[EB/OL]. [2025-01-26].
[25]
Claude. Privacy & legal[EB/OL]. [2025-01-26].
[26]
百度. 文心一言用户协议[EB/OL]. [2025-01-26].
[27]
百度. 文心一言个人信息保护规则[EB/OL]. [2025-01-26].
[28]
百度. 智能体功能服务条款[EB/OL]. [2025-01-26].
[29]
MITCHELL M, WU S, ZALDIVAR A, et al. Model cards for model reporting[J/OL].arXiv: 1810.03993, 2018.
[30]
PUBLIO G C, ESTEVES D, ŁAWRYNOWICZ A, et al. ML-Schema: Exposing the semantics of machine learning with schemas and ontologies[J/OL]. arXiv: 1807.05351, 2018.
[31]
IBM. FactSheet examples overview[EB/OL]. [2025-01-26].
[32]
Amazon. Resources that promote AI transparency[EB/OL]. [2025-01-26].
[33]
OpenAI. OpenAI o1 system card[EB/OL]. [2025-01-26].
[34]
HuggingFace. huggingface_hub[EB/OL]. [2025-01-26].
[35]
BRUCE T R, HILLMANN D I. The continuum of metadata quality: Defining, expressing, exploiting[M]. Chicago: American Library Association, 2004.
[36]
黄莺, 李建阳. 元数据质量评估方法及模型研究[J]. 图书馆学研究, 2013(12): 52-56, 51.
HUANG Y, LI J Y. The study of metadata quality assessment and its model[J]. Research on library science, 2013(12): 52-56, 51.
[37]
程颖. 数字资源元数据质量管理的研究与探索[J]. 图书馆, 2015(7): 66-69, 104.
CHENG Y. Research and practice on the metadata quality management of digital resources[J]. Library, 2015(7): 66-69, 104.
[38]
PARK J R. Metadata quality in digital repositories: A survey of the current state of the art[J]. Cataloging & classification quarterly, 2009, 47(3/4): 213-228.
[39]
OCHOA X, DUVAL E. Automatic evaluation of metadata quality in digital repositories[J]. International journal on digital libraries, 2009, 10(2): 67-91.
[40]
ALJUMAILI M, KARIM R, TRETTEN P. Metadata-based data quality assessment[J]. VINE journal of information and knowledge management systems, 2016, 46(2): 232-250.
[41]
ŠLIBAR B, OREŠKI D, BEGIČEVIĆ REĐEP N. Importance of the open data assessment: An insight into the (meta) data quality dimensions[J]. Sage open, 2021, 11(2): 21582440211023178.
[42]
STVILIA B, HINNANT C C, WU S H, et al. Research project tasks, data, and perceptions of data quality in a condensed matter physics community[J]. Journal of the association for information science and technology, 2015, 66(2): 246-263.
[43]
管珍珍. 信息生命周期视角下智慧情报赋能医药企业核心竞争力研究[D]. 哈尔滨: 黑龙江大学, 2024.
GUAN Z Z. Research on the core competitiveness of pharmaceutical enterprises enabled by intelligent intelligence from the perspective of information lifecycle[D]. Harbin: Helongjiang University, 2024.
[44]
尹文武. 信息生命周期理论下的移动图书馆信息服务质量控制[J]. 图书馆理论与实践, 2017(4): 91-93.
YIN W W. Quality control of mobile library information service based on information life cycle theory[J]. Library theory and practice, 2017(4): 91-93.
[45]
YANG W L, FU R, AMIN M B, et al. The impact of modern AI in metadata management[J]. Human-centric intelligent systems, 2025, 5(3): 323-350.
[46]
RHEE H L. A new lifecycle model enabling optimal digital curation[J]. Journal of librarianship and information science, 2024, 56(1): 241-266.
[47]
KOWALCZYK S T. Modelling the research data lifecycle[J]. International journal of digital curation, 2017, 12(2): 331-361.
[48]
郭海玲, 刘仲山. GDPR对我国跨境数字贸易企业个人数据保护研究: 基于数据生命周期理论[J]. 情报杂志, 2023, 42(10): 194-201.
GUO H L, LIU Z S. The enlightenment of GDPR on personal data protection of Chinese cross-border digital trade enterprises: Based on data life cycle theory[J]. Journal of intelligence, 2023, 42(10): 194-201.
[49]
马静. 数据生命周期视角下高校科学数据共享研究[D]. 哈尔滨: 黑龙江大学, 2024.
MA J. Research on scientific data sharing in universities under the perspective of data lifecycle[D]. Harbin: Helongjiang University, 2024.
[50]
FAIR digital object framework documentation[EB/OL]. [2025-01-26].
[51]
AI智能大模型: 未来发展趋势人工智能AI大模型智能AI发展开发设计[EB/OL]. [2025-01-26].
[1] SHEN Hongjie, SHEN Hongwei, WANG Junli. Generative AI Empowering Information Literacy Education in Digital Libraries: Path Exploration, Challenge Analysis, and Response Strategies [J]. Journal of library and information science in agriculture, 2025, 37(7): 50-60.
[2] ZHANG Tao, WU Sihang. How Achievement Goal Orientation Influences College Students' Usage Behaviors of AI Tutoring Tools: An Empirical Study Based on Dual Mediation [J]. Journal of library and information science in agriculture, 2025, 37(7): 91-105.
[3] JIANG Yumeng. An Investigation and Enlightenment on the Current Status of AI Literacy Education in Nordic University Libraries [J]. Journal of library and information science in agriculture, 2025, 37(7): 19-34.
[4] ZHAI Jun, MENG Zihan, LI Fangsu, SHEN Lixin. AI Guides in Research Libraries of North America under the AI4S Context: Based on the Survey of 125 ARL Libraries [J]. Journal of library and information science in agriculture, 2025, 37(7): 35-49.
[5] XIAO Yufan, CHEN Rui, HUANG Ying. Exploring University Roles in the Innovation Value Chain: A Comparative Study of China's C9 League and the U.S. Ivy League [J]. Journal of library and information science in agriculture, 2025, 37(6): 87-101.
[6] DAI Xinwei, LI Feng. Enlightenment of AI Literacy Educational Designs and Practices at Japanese MDASH Literacy-level Approved Universities [J]. Journal of library and information science in agriculture, 2025, 37(5): 86-101.
[7] SHI Xujie, YUAN Fan, LI Jia. Searching as Learning in the Context of Generative Artificial Intelligence: Technological Pathways, Behavioral Evolution, and Ethical Challenges [J]. Journal of library and information science in agriculture, 2025, 37(5): 40-57.
[8] ZHANG Weichong, XU Chen, ZHU Yiran. Cultivation Path of AI Literacy for Grassroots Civil Servants Based on the Integrated TAM-IDT Model [J]. Journal of library and information science in agriculture, 2025, 37(5): 72-85.
[9] CHEN Jiayong, GONG Jiaoteng, WANG Yuyi. Research of Interdisciplinary Comparison and Collaborative Paradigm on the Concept of Agent in Library Science [J]. Journal of library and information science in agriculture, 2025, 37(5): 27-39.
[10] QIAN Li, WANG Qianying, LIU Yi, ZHANG Yuanzhe, CHANG Zhijun. Agent Technology and Its Applications in Scientific Research [J]. Journal of library and information science in agriculture, 2025, 37(5): 5-14.
[11] SHEN Mengcheng, CHEN Xiuping. Analysis of the Evaluation and Development Pathways for Rural Cultural-Tourism Integration Based on Online Text Data: A Case Study of 26 Mountainous Counties in Zhejiang Province [J]. Journal of library and information science in agriculture, 2025, 37(4): 66-82.
[12] XIAO Keyi, LI Yunfan, CHEN Yingying, PENG Xi. User Experience of Public Library Intelligent Service from the Perspective of Actor Network Theory [J]. Journal of library and information science in agriculture, 2025, 37(4): 51-65.
[13] ZHANG Tao, LYU Qianhui. Generative AI Governance Practices in Europe and the United States and the Enlightenment for China [J]. Journal of library and information science in agriculture, 2025, 37(4): 12-23.
[14] TAN Miao, DAI Mengfei. Empowered Digital Reading Promotion of Historical Documents with Generative AI [J]. Journal of library and information science in agriculture, 2025, 37(4): 83-93.
[15] SHI Zhongyan, LEI Jie, SUN Tan, ZHAO Ruixue, LI Jiao, HUANG Yongwen, XIAN Guojian. Research on DeepSeek-Empowered Low-Cost Construction of Domain-Specific Knowledge Graphs [J]. Journal of library and information science in agriculture, 2025, 37(3): 4-17.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
Editorial Board of Journal of library and information science in agriculture
Tel：(010)82109667　Fax：(010)82109667　E-mail：xuekan＠caas.cn
Powered by Beijing Magtech Co., Ltd.
知识共享许可协议