Qualcomm and Google Cloud to Collaborate on Neural Architecture Search (NAS). With this, the companies want to make the development of AI models more automated instead of manual.
The partnership powers Qualcomm Snapdragon chipsets with Google Cloud Vertex AI NAS technology. This creates the first SoCs that deliver NAS for developers who want to develop and deploy as many automated AI models as possible.
Google Vertex AI NAS is a platform from the cloud giant for developing, deploying and maintaining AI models. Training models developed with the technology require up to 80 percent less code compared to other platforms, according to Google Cloud. The platform would also support various services internally at Google, such as computer vision and implementations for language and structured data.
More AI at the edge
The collaboration should mainly provide more AI functionality at the edge, according to the chip manufacturer. Think of high-quality AI solutions with a low latency that are suitable for devices with a low energy consumption and a regulated memory capacity. Examples are IoT, medical imaging devices, vehicles and mobile devices. The Google Vertex AI NAS technology should also ensure that Qualcomm itself can develop and optimize new AI models for its own applications in a shorter time.
availabilty
The Google Vertex AI NAS technology is integrated into the Qualcomm Neural Processing SDK and runs on the Qualcomm AI Engine. The platform will be available first on the Snapdragon 8, Gen 1 Mobile Platform. A later rollout will take place to the rest of the Snapdragon portfolio for mobile, IoT, automotive and XR platforms, among others.
