Microsoft is a Chief on this yr’s Gartner® Magic Quadrant™ for Information Science and Machine Studying Platforms. Azure AI gives a strong, versatile end-to-end platform for accelerating information science and machine studying innovation.
Microsoft is a Chief on this yr’s Gartner® Magic Quadrant™ for Information Science and Machine Studying Platforms. Azure AI gives a strong, versatile end-to-end platform for accelerating information science and machine studying innovation whereas offering the enterprise governance that each group wants within the period of AI.
In Could 2024, Microsoft was additionally named a Chief for the fifth yr in a row within the Gartner® Magic Quadrant™ for Cloud AI Developer Companies, the place we positioned furthest for our Completeness of Imaginative and prescient. We’re happy by these recognitions from Gartner as we proceed serving to prospects, from giant enterprises to agile startups, convey their AI and machine studying fashions and purposes into manufacturing securely and at scale.
Azure AI is on the forefront of purpose-built AI infrastructure, accountable AI tooling, and serving to cross-functional groups collaborate successfully utilizing Machine Studying Operations (MLOps) for generative AI and conventional machine studying tasks. Azure Machine Studying gives entry to a broad collection of basis fashions within the Azure AI mannequin catalog—together with the current releases of Phi-3, JAIS, and GPT-4o—and instruments to fine-tune or construct your personal machine studying fashions. Moreover, the platform helps a wealthy library of open-source frameworks, instruments, and algorithms in order that information science and machine studying groups can innovate in their very own manner, all on a trusted basis.
Speed up time to worth with Azure AI infrastructure
“We’re now capable of get a functioning mannequin with related insights up and working in simply a few weeks due to Azure Machine Studying. We’ve even managed to supply verified fashions in simply 4 to 6 weeks.”
Dr. Nico Wintergerst, Employees AI Analysis Engineer at relayr GmbH
Azure Machine Studying helps organizations construct, deploy, and handle high-quality AI options rapidly and effectively, whether or not constructing giant fashions from scratch, working inference on pre-trained fashions, consuming fashions as a service, or fine-tuning fashions for particular domains. Azure Machine Studying runs on the identical highly effective AI infrastructure that powers a few of the world’s hottest AI providers, reminiscent of ChatGPT, Bing, and Azure OpenAI Service. Moreover, Azure Machine Studying’s compatibility with ONNX Runtime and DeepSpeed can assist prospects additional optimize coaching and inference time for efficiency, scalability, and energy effectivity.
Whether or not your group is coaching a deep studying mannequin from scratch utilizing open supply frameworks or bringing an current mannequin into the cloud, Azure Machine Studying allows information science groups to scale out coaching jobs utilizing elastic cloud compute assets and seamlessly transition from coaching to deployment. With managed on-line endpoints, prospects can deploy fashions throughout highly effective CPU and graphics processing unit (GPU) machines without having to handle the underlying infrastructure—saving effort and time. Equally, prospects don’t have to provision or handle infrastructure when deploying basis fashions as a service from the Azure AI mannequin catalog. This implies prospects can simply deploy and handle hundreds of fashions throughout manufacturing environments—from on-premises to the sting—for batch and real-time predictions.
Streamline operations with versatile MLOps and LLMOps
“Immediate movement helped streamline our improvement and testing cycles, which established the groundedness we required for ensuring the client and the answer have been interacting in a practical manner.”
Fabon Dzogang, Senior Machine Studying Scientist at ASOS
Machine studying operations (MLOps) and giant language mannequin operations (LLMOps) sit on the intersection of individuals, processes, and platforms. As information science tasks scale and purposes change into extra advanced, efficient automation and collaboration instruments change into important for reaching high-quality, repeatable outcomes.
Azure Machine Studying is a versatile MLOps platform, constructed to help information science groups of any measurement. The platform makes it straightforward for groups to share and govern machine studying belongings, construct repeatable pipelines utilizing built-in interoperability with Azure DevOps and GitHub Actions, and constantly monitor mannequin efficiency in manufacturing. Information connectors with Microsoft sources reminiscent of Microsoft Cloth and exterior sources reminiscent of Snowflake and Amazon S3, additional simplify MLOps. Interoperability with MLflow additionally makes it seamless for information scientists to scale current workloads from native execution to the cloud and edge, whereas storing all MLflow experiments, run metrics, parameters, and mannequin artifacts in a centralized workspace.
Azure Machine Studying immediate movement helps streamline the whole improvement cycle for generative AI purposes with its LLMOps capabilities, orchestrating executable flows comprised of fashions, prompts, APIs, Python code, and instruments for vector database lookup and content material filtering. Azure AI immediate movement can be utilized along with widespread open-source frameworks like LangChain and Semantic Kernel, enabling builders to convey experimental flows into immediate movement to scale these experiments and run complete evaluations. Builders can debug, share, and iterate on purposes collaboratively, integrating built-in testing, tracing, and analysis instruments into their CI/CD system to repeatedly reassess the standard and security of their utility. Then, builders can deploy purposes when prepared with one click on and monitor flows for key metrics reminiscent of latency, token utilization, and technology high quality in manufacturing. The result’s end-to-end observability and steady enchancment.
Develop extra reliable fashions and apps
“The accountable AI dashboard gives useful insights into the efficiency and habits of laptop imaginative and prescient fashions, offering a greater degree of understanding into why some fashions carry out in a different way than others, and insights into how numerous underlying algorithms or parameters affect efficiency. The profit is better-performing fashions, enabled and optimized with much less effort and time.”
—Teague Maxfield, Senior Supervisor at Constellation Clearsight
AI ideas reminiscent of equity, security, and transparency will not be self-executing. That’s why Azure Machine Studying gives information scientists and builders with sensible instruments to operationalize accountable AI proper of their movement of labor, whether or not they should assess and debug a conventional machine studying mannequin for bias, defend a basis mannequin from immediate injection assaults, or monitor mannequin accuracy, high quality, and security in manufacturing.
The Accountable AI dashboard helps information scientists assess and debug conventional machine studying fashions for equity, accuracy, and explainability all through the machine studying lifecycle. Customers may also generate a Accountable AI scorecard to doc and share mannequin efficiency particulars with enterprise stakeholders, for extra knowledgeable decision-making. Equally, builders in Azure Machine Studying can evaluation mannequin playing cards and benchmarks and carry out their very own evaluations to pick out the very best basis mannequin for his or her use case from the Azure AI mannequin catalog. Then they will apply a defense-in-depth strategy to mitigating AI dangers utilizing built-in capabilities for content material filtering, grounding on contemporary information, and immediate engineering with security system messages. Analysis instruments in immediate movement allow builders to iteratively measure, enhance, and doc the influence of their mitigations at scale, utilizing built-in metrics and customized metrics. That manner, information science groups can deploy options with confidence whereas offering transparency for enterprise stakeholders.
Learn extra on Accountable AI with Azure.
Ship enterprise safety, privateness, and compliance
“We wanted to decide on a platform that offered best-in-class safety and compliance as a result of delicate information we require and one which additionally provided best-in-class providers as we didn’t wish to be an infrastructure internet hosting firm. We selected Azure due to its scalability, safety, and the immense help it affords by way of infrastructure administration.”
—Michael Calvin, Chief Technical Officer at Kinectify
In right now’s data-driven world, efficient information safety, governance, and privateness require each group to have a complete understanding of their information and AI and machine studying programs. AI governance additionally requires efficient collaboration between various stakeholders, reminiscent of IT directors, AI and machine studying engineers, information scientists, and threat and compliance roles. Along with enabling enterprise observability via MLOps and LLMOps, Azure Machine Studying helps organizations be certain that information and fashions are protected and compliant with the very best requirements of safety and privateness.
With Azure Machine Studying, IT directors can limit entry to assets and operations by consumer account or teams, management incoming and outgoing community communications, encrypt information each in transit and at relaxation, scan for vulnerabilities, and centrally handle and audit configuration insurance policies via Azure Coverage. Information governance groups may also join Azure Machine Studying to Microsoft Purview, in order that metadata on AI belongings—together with fashions, datasets, and jobs—is robotically revealed to the Microsoft Purview Information Map. This allows information scientists and information engineers to watch how parts are shared and reused and study the lineage and transformations of coaching information to know the influence of any points in dependencies. Likewise, threat and compliance professionals can observe what information is used to coach fashions, how base fashions are fine-tuned or prolonged, and the place fashions are employed throughout totally different manufacturing purposes, and use this as proof in compliance studies and audits.
Lastly, with the Azure Machine Studying Kubernetes extension enabled by Azure Arc, organizations can run machine studying workloads on any Kubernetes clusters, making certain information residency, safety, and privateness compliance throughout hybrid public clouds and on-premises environments. This enables organizations to course of information the place it resides, assembly stringent regulatory necessities whereas sustaining flexibility and management over their MLOps. Prospects utilizing federated studying methods together with Azure Machine Studying and Azure confidential computing may also practice highly effective fashions on disparate information sources, all with out copying or shifting information from safe places.
Get began with Azure Machine Studying
Machine studying continues to remodel the best way companies function and compete within the digital period—whether or not you wish to optimize your small business operations, improve buyer experiences, or innovate. Azure Machine Studying gives a strong, versatile machine studying and information science platform to operationalize AI innovation responsibly.
*Gartner, Magic Quadrant for Information Science and Machine Studying Platforms, By Afraz Jaffri, Aura Popa, Peter Krensky, Jim Hare, Raghvender Bhati, Maryam Hassanlou, Tong Zhang, 17 June 2024.
Gartner, Magic Quadrant for Cloud AI Developer Companies, Jim Scheibmeir, Arun Batchu, Mike Fang, Printed 29 April 2024.
GARTNER is a registered trademark and repair mark of Gartner, Inc. and/or its associates within the U.S. and internationally, Magic Quadrant is a registered trademark of Gartner, Inc. and/or its associates and is used herein with permission. All rights reserved.
Gartner doesn’t endorse any vendor, services or products depicted in its analysis publications and doesn’t advise expertise customers to pick out solely these distributors with the very best scores or different designation. Gartner analysis publications include the opinions of Gartner’s Analysis & Advisory group and shouldn’t be construed as statements of truth. Gartner disclaims all warranties, expressed or implied, with respect to this analysis, together with any warranties of merchantability or health for a specific function.
This graphic was revealed by Gartner, Inc. as half of a bigger analysis doc and needs to be evaluated within the context of the whole doc. The Gartner doc is accessible upon request from this hyperlink.