|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/isoforartificialintelligence/
课程评论：没有评论
课程名称：人工智能的ISO标准：ISO 42001及其后续发展 课程概述： ISO 42001是第一个定义可认证的人工智能管理系统的ISO标准。作为一个可认证的人工智能标准，它为组织提供了一种在人工智能应用中建立可信度的标志。这样可以促进商业规模的扩展，这对于推动增长和扩张至关重要，因为全球的组织都在寻求通过ISO等知名标准来提升人工智能的可信度。ISO 42001专注于在组织层面上负责任地管理人工智能，提供实施有效的人工智能治理和管理系统的框架。 本课程将涵盖人工智能的基本概念以及与人工智能相关的ISO标准。课程由一位在大型跨国公司及中小企业拥有丰富人工智能部署经验的IT资深人士授课。虽然ISO 42001是我们关注的主要可认证标准，但我们还将讨论其他相关标准，因为它们涉及人工智能的重要方面并同样关键。 相关标准包括： - ISO 23053：提供一个结构化框架，用于构建和维护以机器学习为中心的人工智能系统，确保其生命周期中的一致性和可靠性。 - ISO 5338：提供人工智能系统的整体生命周期模型，补充ISO 23053，涵盖更广泛的人工智能场景，包括非机器学习方法和更长的运营阶段。 - ISO 38507：关注人工智能治理、数据治理和风险管理，为道德和负责任的人工智能系统部署提供基础。 - ISO TR 24027：专注于识别、评估和缓解人工智能系统中的偏见，促进人工智能结果的公平性和公正性。 - ISO TR 24028：探讨人工智能的可信度，涉及透明度、可靠性和伦理等原则，以增强用户信心。 - ISO 24029：提供评估神经网络鲁棒性的指导，确保它们在不同条件下，包括对抗性场景下的可靠性能。 这些标准共同构成了一个全面的生态系统，旨在以负责任的方式开发、部署和管理人工智能系统。
ISO 42001 is the first ISO standard defining a certifiable management system for AI. As a certifiable AI standard, it helps provide organisations with a mark of trustworthiness in AI use. This can help promote commercial scalability - vital for encouraging growth and expansion, because organization around the world are looking at making AI trustworthy through reputed standards like ISO. ISO 42001 focuses on managing AI responsibly at an organizational level, providing a framework for implementing effective AI governance and management systems.In this course, we'll cover both AI concepts and the ISO standards relevant to AI. This course is taught by an IT veteran who has rich experience in AI deployments in large MNCs and Small and medium businesses.While ISO 42001 is the primary certifiable standard and our focus area, we cover other related standards since they address critical aspects of AI and are equally important.ISO 23053: Offers a structured framework for building and maintaining ML-centric AI systems, ensuring consistency and reliability throughout their lifecycle.ISO 5338: Provides a holistic lifecycle model for AI systems, complementing ISO 23053 by covering broader AI scenarios, including non-ML methods and extended operational phases.ISO 38507: Addresses AI governance, data governance, and risk management, offering a foundation for ethical and responsible AI system deployment.ISO TR 24027: Focuses on identifying, assessing, and mitigating bias in AI systems, promoting fairness and equity in AI outcomes. TR stands for technical report and we will explain that when we cover bias.ISO TR 24028: Explores trustworthiness in AI, addressing principles like transparency, reliability, and ethical considerations to build user confidence.ISO 24029: Provides guidance for evaluating the robustness of neural networks, ensuring they perform reliably under various conditions, including adversarial scenarios.Together, these standards form a comprehensive ecosystem for developing, deploying, and managing AI systems responsibly.
