当前位置:首页 > 报告详情

人工智能未来的主要挑战.pdf

上传人: 哆哆 编号:631232 2025-04-19 21页 2.98MB

1、AI future main challengesJuliette MATTIOLI16?AI Main Technology Trends(mainly from 2024 AI Gartner hype curve)?Computer VisionCloud AI ServicesEdge AIKnowledge GraphSovereign AIMLOps/AI OpsSynthetic DataAI Engineering?Quantum AIMNeuro-symbolic AI SPrompt EngineeringResponsible AIFoundation ModelsFAI

2、 GovernanceDecision IntelligencePhysics/Geometry Informed NNAI TRiSM*AI TRiSM*:tackling Trust,Risk and Security in AI Model Federated MLReinforcement LearningMulti-agent SystemsCausal AINeuromorphic ComputingGenerative AIGLarge Language ModelL17?AI GovernanceAI Main Technology Trends:Trustworthy and

3、 Responsible AI?Computer VisionCloud AI ServicesEdge AIKnowledge GraphNeuromorphic ComputingMLOps/AI OpsSynthetic DataAI Engineering?Quantum AIMNeuro-symbolic AI SPrompt EngineeringQResponsible AIFoundation ModelsF FDecision IntelligencePhysics/Geometry Informed NNGenerative AIGAI TRiSM*AI TRiSM*:ta

4、ckling Trust,Risk and Security in AI Model Federated MLReinforcement LearningMulti-agent SystemsCausal AISovereign AITrustworthy and Responsible AILarge Language ModelL18?Trustworthy and Responsible AIValidityTo guaranty that an AI-based system will do what it is meant to do,all what it is meant to

5、do and only what is meant to doExplainabilityTo be able to provide human-level,understandable and context-relevant justifications and explanationsSecurityTo ensure robustness and resilience to adversarial conditions,such as decoying and cyber-attacksResponsibilityTo be compliant with ethical,legal a

6、nd regulatory frameworksThales TrUEAI Transparent,Understandable,Ethical19?Data&AI RegulationGDPRApplicable since May 2018Ensure that companies handle personal data in a responsible and accountable way,and that individuals have greater control over their dataEU DATA GOVERNANCE ACTApplicable Septembe

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
本文主要介绍了AI技术的发展趋势和挑战,以及如何实现可信赖和负责任的AI。文章指出,AI技术的发展趋势包括计算机视觉、云AI服务、边缘AI、知识图谱、主权AI、MLOps/AI Ops、合成数据、AI工程等。同时,AI技术面临的挑战包括可解释性、安全性、责任性等。为了实现可信赖和负责任的AI,文章提出了透明度和可解释性、隐私和数据保护、鲁棒性和安全性、问责制、以人为中心的设计、社会和环境影响、合作和多方利益相关者参与等关键原则。此外,文章还介绍了AI TRiSM(AI信任、风险和安全管理)的概念,以及如何通过MLOps工具链来支持数据驱动的AI部署。总的来说,文章强调了在AI技术的发展过程中,需要关注可信赖性和负责任性,以确保AI技术能够为社会带来积极的影响。
如何在AI中实现可解释性和透明度? 如何确保AI系统的安全性和鲁棒性? 如何通过AI实现社会和环境影响的积极贡献?
客服
商务合作
小程序
服务号
折叠