Intel and Microsoft work together to promote large-scale deployment of AI and leverage the potential of edge-cloud collaboration

Today, Intel and Microsoft jointly held the 2022 China AI Developer Summit online. Matthew Formica, senior director of Intel OpenVINO ^™ developer ecosystem, Qiu Lili, deputy director of Microsoft Research Asia, and industry technical experts from Intel, Intel AI Developer Community, Microsoft Artificial Intelligence and Internet of Things Lab, Microsoft, EdgeX Foundry Technology Community, Advantech Technology and Hanshuo Technology attended the summit and explained to everyone the deep understanding, key technologies and practical experience of efficient development and large-scale deployment of AI solutions.

Driven by new infrastructure such as 5G, big data centers, and AI, the digitalization level of enterprises, the complexity and professionalism of their businesses have all been significantly improved. This has put forward new requirements for enterprises to optimize the allocation of computing power, data timeliness, system stability, and flexibility across work scenarios when deploying in the cloud. Therefore, edge-cloud collaboration has become an important trend in future development, and many companies have begun to build edge-cloud collaboration models and systems to collaboratively solve computing problems under the development of big data.

As these companies transform, large-scale deployment of AI solutions has become a top priority.

When choosing hardware, professionals should consider reducing latency, improving performance per watt, and optimizing costs as much as possible, and Intel processors can fully meet these needs. At the same time, Intel can also provide developers with a full-stack AI software product portfolio. This product portfolio is first of all oneAPI as the basic layer from the bottom up, which can support a series of Intel processors with general functions. Secondly ^, for inference optimization and deployment, OpenVINO ^™ can be adjusted for each Intel processor , helping developers to easily achieve scalable and high-performance deep learning inference. In addition, Intel also promotes collaboration across the ecosystem and helps partners grow and innovate by launching Intel® Market Ready Solutions .

At the summit, Intel technical experts also detailed how to use OpenVINO ^™ on Azure to quickly achieve high-performance AI reasoning computing ^{. Since the launch of OpenVINO ™} in 2018 , Intel has helped hundreds of thousands of developers significantly improve AI reasoning performance and expand its applications from edge computing to enterprises and clients. At the end of the first quarter of 2022, Intel launched a new version of the OpenVINO ^™ toolkit, OpenVINO ^™ 2022.1 . The new features are mainly developed based on developer feedback over the past three and a half years, including more deep learning frameworks and model support, more device portability options, and higher reasoning performance and fewer code changes, designed to simplify and automate optimization work. The OpenVINO ^™ 2022.1 version adds hardware auto-discovery and auto-optimization capabilities, allowing software developers to achieve excellent performance on any platform. In addition, Intel's Open Model Zoo also contains more than 270 pre-trained optimization models, including object detection, action recognition, machine question and answer, and speech-to-text. By providing these models to developers, the model design steps of developers can be greatly simplified.

In addition, industry experts from the EdgeX Foundry technology community also explained how to build scalable AI/IoT solutions based on Intel architecture and with the help of EdgeX and Azure to accelerate the transformation of edge data into effective insights, thereby achieving edge-cloud collaboration. Finally, technical experts from Hanshuo Technology shared the company's smart retail AI solution built on Intel architecture , taking the retail industry as an example to fully demonstrate the important role played by AIoT in driving the digital transformation of vertical industries. This solution combines Intel's full-stack technology optimization from software to hardware and Hanshuo Technology's long-term technological innovation and industry best practices in the field of "retail + AI". It can help retail customers build a high-performance, high-insight, high-economic, and easy-to-implement smart retail management system, thereby helping to solve the thorny problems in shelf merchandise management, self-service checkout, and merchandise loss prevention, bringing greater value to global retail customers.

As all industries are actively undergoing digital transformation, enterprises are increasingly demanding the transmission and processing speed of massive amounts of data of different types. Therefore, in the process of achieving edge-cloud collaboration, the importance of the network is self-evident.

Today, AI solutions with high social and commercial value, such as smart retail and autonomous driving, are profoundly reshaping the entire society. Intel will work with a wide range of ecosystem partners to continuously empower developers by continuously launching more optimized new products and tools, so that they can collaborate to complete the efficient training and high-performance deployment of AI models at home or in the office, and achieve end-to-end machine learning full life cycle management. At the same time, it will simplify the development and operation of high-value AI solutions, promote the large-scale deployment of high-value AI solutions in the world, and promote the collaborative innovation and development of edge-cloud.