Huawei sees progress in AI computing
Chinese tech heavyweight Huawei Technologies Co is making solid progress in building a commercial ecosystem for the application of its artificial intelligence computing technologies.
This progress comes as AI is gradually penetrating all aspects of people's work and lives, with its application scope widened from the internet sector to thousands of industries to help boost operational efficiency and reduce costs.
According to Huawei, based on its Ascend series of AI processors, the Huawei Atlas AI computing solution offers a broad portfolio of products, including modules, cards, servers and clusters. The solution enables all-scenario AI infrastructure through the device-edge-cloud process, covering full-pipeline inference and training for AI deep learning.
The technology and commercial ecosystem for Huawei's Ascend AI chips is starting to take shape. So far, Ascend AI has teamed up with more than 500 partners to jointly incubate more than 900 industrial solutions. The number of third-partly developers for Ascend AI has exceeded 700,000, with Ascend-related curricula covering more than 100,000 students in 72 colleges and universities.
"AI has three key elements, namely computing power, algorithms and data, which constitutes the foundation of AI," an AI computing expert from Huawei who declined to be named said.
According to the expert, Huawei's self-developed AI computing framework MindSpore is critical to making AI application development easier, AI applications more pervasive and accessible, and ensuring privacy protection.
The moves represent crucial efforts by Huawei to build a full, all-scenario AI portfolio. They are also designed to beef up Huawei's in-house research and development capabilities, as Washington's ban on the company continues.
According to Huawei, its Ascend AI can not only help big companies but also small and medium-sized enterprises embrace AI, as it can help solve the problems of difficult and expensive computing power in traditional computing methods.