跳过导航

Azure 机器学习

通过试验和模型管理,使用端到端、可缩放的受信平台,让所有人都能够获得 AI 体验

加速端到端机器学习生命周期

Empower data scientists and developers with a wide range of productive experiences to build, train, and deploy machine learning models and foster team collaboration. Accelerate time to market with industry-leading MLOps—machine learning operations, or DevOps for machine learning. Innovate on a secure, trusted platform, designed for responsible machine learning.

适合所有技能水平的机器学习

Productivity for all skill levels, with Jupyter Notebooks, drag-and-drop designer, and automated machine learning

端到端 MLOps

Robust MLOps capabilities that enable creation and deployments of models at scale using automated and reproducible machine learning workflows

负责任的机器学习创新

Rich set of built-in responsible capabilities to understand, protect, and control data, models, and processes

开放且可交互

Best-in-class support for open-source frameworks and languages including MLflow, Kubeflow, ONNX, PyTorch, TensorFlow, Python, and R

Boost productivity with machine learning for all skill levels

使用满足各种技能水平需求的工具,快速构建并部署机器学习模型。结合使用内置的 Jupyter 笔记本与 Intellisense 或拖放式设计器。使用自动化机器学习加速模型创建,并访问强大的特征工程、算法选择和超参数清除功能。通过共享的数据集、笔记本、模型和可自定义的仪表板(可跟踪机器学习过程的各个方面)提高团队效率。

使用 MLOps 执行大规模操作

Take advantage of MLOps to streamline the machine learning lifecycle, from building models to deployment and management. Create reproducible workflows with machine learning pipelines, and train, validate, and deploy thousands of models at scale, from the cloud to the edge. Use managed online and batch endpoints to seamlessly deploy and score models without managing the underlying infrastructure. Use Azure DevOps or GitHub Actions to schedule, manage, and automate the machine learning pipelines, and use advanced data-drift analysis to improve model performance over time.

构建负责任的机器学习解决方案

Access state-of-the-art responsible machine learning capabilities to understand, control, and help protect your data, models, and processes. Explain model behavior during training and inferencing, and build for fairness by detecting and mitigating model bias. Preserve data privacy throughout the machine learning lifecycle with differential privacy techniques and use confidential computing to secure machine learning assets. Automatically maintain audit trails, track lineage, and use model datasheets to enable accountability.

在灵活的开放平台上创新

Get built-in support for open-source tools and frameworks for machine learning model training and inferencing. Use familiar frameworks like PyTorch, TensorFlow, or scikit-learn, or the open and interoperable ONNX format. Choose the development tools that best meet your needs, including popular IDEs, Visual Studio Code, Jupyter Notebooks, and CLIs, or languages such as Python and R. Use ONNX Runtime to optimize and accelerate inferencing across cloud and edge devices. Track all aspects of your training experiments using MLflow.

通过 Azure 培养机器学习技能

了解有关 Azure 上的机器学习的详细信息,并参与本次为期 30 天的学习旅程的实践教程。在本次学习旅程结束时,你就为参加 Azure 数据科学家助理认证做好了准备。

先进的安全、治理和混合基础结构

  • 在混合基础结构中,借助 Azure Arc 互操作性在本地、多云环境和边缘的 Kubernetes 群集上训练模型。
  • 访问安全功能,例如基于角色的访问、自定义机器学习角色、虚拟网络和专用链接。使用策略、审核线索、配额和成本管理来管理治理。
  • 通过 60 项认证(包括 FedRAMP High 和 DISA IL5)的综合产品组合,优化合规性。

主要服务功能

协作式笔记本

Maximize productivity with IntelliSense, easy compute and kernel switching, and offline notebook editing. Launch your notebook in Visual Studio Code for a rich development experience, including secure debugging and support for Git source control.

自动执行机器学习

快速创建用于分类、回归和时序预测的准确模型。使用模型可解释性来了解构建模型的方式。

拖放式机器学习

使用机器学习工具(如设计器)和模块进行数据转换、模型训练和评估,或者轻松地创建和发布机器学习管道。

数据标签

快速准备数据、管理和监视标记项目,并借助机器学习辅助标记自动执行迭代任务。

MLOps

Use the central registry to store and track data, models, and metadata. Automatically capture lineage and governance data. Use Git to track work and GitHub Actions to implement workflows. Manage and monitor runs, or compare multiple runs for training and experimentation. Use managed endpoints to operationalize model deployment and scoring, log metrics, and perform safe model rollouts.

自动缩放计算

使用托管计算来分发训练并快速测试、验证和部署模型。在工作区内共享 CPU 和 GPU 群集并自动缩放,以满足机器学习需求。

与其他 Azure 服务深度集成

借助与 Microsoft Power BI 和 Azure 服务(例如 Azure Synapse Analytics、Azure 认知搜索、Azure 数据工厂、Azure Data Lake、Azure Arc 和 Azure Databricks)的内置集成来提高工作效率。

混合和多云支持

Run machine learning on existing Kubernetes clusters on-premises, in multicloud environments, and at the edge with Azure Arc. Use the one-click machine learning agent to start training models more securely, wherever your data lives.

强化学习

将强化学习扩展到强大的计算群集,支持多代理方案并访问开源强化学习算法、框架以及环境。

负责任的机器学习

Get model transparency at training and inferencing with interpretability capabilities. Assess model fairness through disparity metrics and mitigate unfairness. Help protect data with differential privacy and confidential machine learning pipelines.

企业级安全

使用网络隔离和专用链接功能、资源和操作基于角色的访问控制、自定义角色以及计算资源的托管标识,更安全地构建和部署模型。

成本管理

使用工作区和资源级别配额限制,更好地管理 Azure 机器学习计算实例的资源分配。

仅对所需资源付费,无前期成本

查看 Azure 机器学习定价

掌握 Azure 机器学习

掌握使用 TensorFlow、Spark 和 Kubernetes 在 Azure 中生成高度可缩放的自动化端到端机器学习模型和管道的专家技术。

数据科学原理

许多从事数据工作的人已具备了数学、编程或域专业知识方面的技能,但对于严格意义上的数据科学而言,这三方面的技能缺一不可。这一综合性的电子书有助于填补这些空白。

Forrester Wave 2020 年度领导者

Forrester 在“The Forrester Wave™:基于笔记本的预测分析和机器学习,2020 年第 3 季度”中将 Microsoft Azure 机器学习评为“领导者”。

如何使用 Azure 机器学习

进入工作室 Web 体验

构建和定型

部署和管理

步骤 1(共 1 步)

创作新模型,并将计算目标、模型、部署、指标和运行历史记录存储在云中。

步骤 1(共 1 步)

使用自动化机器学习功能识别算法和超参数,并在云中跟踪试验。使用笔记本或拖放设计器来创作模型。

步骤 1(共 1 步)

将机器学习模型部署到云端或边缘,监视性能并根据需要重新训练。

立即开始使用 Azure 机器学习

通过注册 Azure 免费帐户获得即时访问权限和 $200 额度。

登录 Azure 门户

客户使用 Azure 机器学习

"The model we deployed on Azure Machine Learning helped us choose three new retail locations. Those stores exceeded their revenue plans by over 200 percent [that] December, the height of our season, and within months of opening were among the best-performing stores in their districts."

Jolie Vitale,Carhartt 公司的 BI 和分析主管
Carhartt

"MLOps is at the core of our product. Because of its reproducible ML pipelines, ... registered models, and automatic model scoring, we're definitely detecting things that we missed before. Which, in terms of risk management, is really, really important."

Ignasi Paredes-Oliva,雀巢全球安全运营中心首席数据科学家
Nestle Italia

"Azure Machine Learning allows us to manage the entire lifecycle, from experimentation and development to production and enhancements."

Joey Chua,AGL 机器学习工程高级经理
AGL

"With model interpretability in Azure Machine Learning, we have a high degree of confidence that our machine learning model is generating meaningful and fair results."

Scandinavian Airlines 数据分析与人工智能负责人 Daniel Engberg
Scandinavian Airlines

"We've used the MLOps capabilities in Azure Machine Learning to simplify the whole machine learning process. That allows us to focus more on data science and let Azure Machine Learning take care of end-to-end operationalization."

Michael Cleavinger, Senior Director of Shopper Insights Data Science and Advanced Analytics, PepsiCo
PepsiCo

"We see Azure Machine Learning and our partnership with Microsoft as critical to driving increased adoption and acceptance of AI from the regulators."

Alex Mohelsky,EY Canada 合作伙伴兼顾问、数据分析与 AI 负责人
EY

Azure 机器学习更新、博客和公告

有关 Azure 机器学习的常见问题

  • 该服务已在一些国家/地区正式发布,即将在其他国家/地区正式发布。
  • Azure 机器学习的服务级别协议 (SLA) 为 99.9% 运行时间。
  • Azure 机器学习工作室是机器学习的首要资源。此功能为数据科学家和开发人员提供了一个集中的场所,他们可使用其中的所有项目来构建、训练和部署机器学习模型。

你随时可以开始设置 Azure 免费帐户