Inside Advanced Scale Challenges|Saturday, March 24, 2018
  • Subscribe to EnterpriseTech Weekly Updates: Subscribe by email

Intel, Microsft Partner on Machine Vision, AI 

Source: Intel Corp.

Intel and Microsoft are reuniting to support the chip maker’s vision processing unit on the software giant’s machine learning platform. The partners said the combination would allow AI inference at the network edge as Intel’s Movidius unit looks to expand its machine vision and AI offerings.

The partners said this week the combination of the Intel Movidius Myriad X vision processor with the Microsoft platform would allow developers to explore machine learning tasks within the Microsoft OS. The Intel VPU is among a list of processors supported by Microsoft for handling AI workloads. The collaboration with Intel will focus on deploying deep neural network applications on Windows clients, the partners said Wednesday (March 7).

Intel (NASDAQ: INTC) bills the Movidius Myriad X vision processor as a system-on-chip with a dedicated neural computing engine for deep learning inference applications at the network edge. The third-generation chip is designed to boost the performance of deep neural networks while handling other specific AI tasks that can slow other system hardware.

The approach aims to provide AI developers with another option for dedicated deep learning inference for training machine learning models, thereby freeing other hardware for other workloads. For edge devices, the vision chip running on the Microsoft machine learning platform also would reduce power consumption and eliminate the need to write custom code.

The partners said potential applications include Windows client applications such as personal assistants, biometric security and advance search capabilities.

Intel unveiled the Myriad X VPU last summer, about a year after acquiring visual processing specialist Movidius. The partnership with Microsoft (NASDAQ: MSFT) would allow the chip maker to scale up VPU manufacturing for AI applications via Microsoft’s huge installed base of Windows clients. The deal is also designed to move AI capabilities like deep learning inference applications from the datacenter to edge devices, noted Remi El-Ouazzane, a company vice president and general manager of Intel Movidius.

Market watchers note that visual processing engines will be critical to emerging AI applications such as autonomous vehicles, drone fleets and Internet of Things (IoT) sensor networks.

Among the goals of the chip maker’s vision processing strategy is moving AI solutions from cloud implementations to the network edge. Hence, Intel and other chip makers like Arm are focusing on low-power IoT applications. While the U.K. chip intellectual property vendor has been focusing on requirements like securing IoT devices, it also introduced a machine learning platform in February designed to boost device functionality with capabilities like object detection.

Meanwhile, Intel’s partnership with Microsoft signals its intent to also deliver new categories of intelligent edge devices.

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).

Add a Comment

Share This