Qualcomm and Google Cloud will collaborate on Neural Architecture Search (NAS). With this, the companies aim to automate the development of AI models.
The collaboration will equip Qualcomm Snapdragon chipsets with Google Cloud’s Vertex AI NAS technology. This creates the first SoCs to deliver Neural Architecture Search (NAS) for developers who want to develop and deploy AI models in an automated fashion.
Google Vertex AI NAS is a platform for developing, deploying and maintaining AI models. According to Google Cloud, training models developed with the technology require up to 80 percent less code than other platforms. The platform also supports various Google services such as computer vision and implementations for language and structured data.
More AI functionality at the edge
Above all, the collaboration between Google and Qualcomm should deliver more AI functionality at the edge. This includes high-performance AI solutions with low latency, suitable for devices with low power consumption and regulated memory capacity. Examples include IoT, devices for medical images, vehicles and mobile devices. The Google Vertex AI NAS technology should also allow Qualcomm to develop and optimize new AI models for its own applications in less time.
The Google Vertex AI NAS technology will be integrated into the Qualcomm Neural Processing SDK and run on the Qualcomm AI Engine. The platform will first be available on the Snapdragon 8 Gen 1 Mobile Platform. A later rollout will take place for the rest of the Snapdragon portfolio for mobile, IoT, automotive and XR platforms.
Tip: Google’s new Vertex AI platform helps companies adopt ML faster