Data-Gouv-FR/bonnes-pratiques-pour-la-mise-en-oeuvre-dune-ia-frugale
收藏资源简介:
该数据集提供了节俭人工智能通用参考框架中制定的最佳实践,共有31条,按以下因素分类:实施产生的节俭收益(低、中、高)、实施所需努力强度(低、中、高)、干预的AI生命周期阶段(横向、初始化、设计与开发、验证与确认、部署、运营与监控、持续验证、重新评估、停用)以及涉及的元素(服务、数据、基础设施)。每条最佳实践包括其在参考框架中的编号、标题、主题、描述、详细实施方法和来源(如见证或参考文献)。数据集旨在帮助组织根据其优先级灵活选择实践,以促进AI项目的环境可持续性。数据格式为CSV(UTF-8编码,分号分隔)和Parquet,源自法国政府与AFNOR的合作项目,基于CC-BY 4.0许可。
This dataset contains 31 best practices developed under the General Reference Framework for Frugal Artificial Intelligence. These practices are classified along four dimensions: 1) Frugal benefits yielded by implementation (low, medium, high); 2) Effort intensity required for implementation (low, medium, high); 3) Targeted AI lifecycle stage of the intervention (cross-cutting, initiation, design and development, verification and validation, deployment, operation and monitoring, continuous validation, re-evaluation, decommissioning); 4) Involved elements (services, data, infrastructure). Each best practice includes its number in the reference framework, title, topic, detailed description, implementation method, and source (e.g., testimonies or references). The dataset is designed to enable organizations to flexibly select practices based on their priorities, thereby advancing the environmental sustainability of AI projects. The dataset is available in CSV (UTF-8 encoded, semicolon-separated) and Parquet formats, originating from a collaborative project between the French Government and AFNOR, and is licensed under CC-BY 4.0.




