What Is Edge AI and How Does It Work? Hi, everyone, my name is Gary. I am a Software product manager in Advantech. AI is increasingly popular. Did you know that AI technologies have improved quality and production of most milk we drink everyday? Service AI robots are adopted in more and more restaurants to provide interactive and precise services. More and more cases use AI to ensure optimal customer experiences. You must hear of ChatGPT and Whisper. They much improve interactivity and customer satisfaction for robotic services. In the future, there will be more numerous, more innovative and more valuable AI technologies coming out. They will encourage more AI and robot applications. However, each AI application always takes lots of efforts and time from engineers. There are so many AI development kits in the market. It took lots of time for engineers to try out, then develop with necessary dev. kits and finally deploy and verify their developed AI in the field. When dealing with such a complex pipeline, there are several pain points to them. How to evaluate and compare AI inference performance at edge quickly? How to retrain a model on demand efficiently and precisely? How to jump into develop AI applications right away with inference-accelerated SDKs? How to manage massive AIoT devices in the field and deploy and maintain edge AI applications efficiently? ADVANTECH has been involved with industrial control domains for more than forty years and is aware of the bottlenecks to customers when they developed their AI applications. To mitigate those bottlenecks, ADVANTECH will launch the new Edge AI Suite. Edge AI Suite is one tool set for customers to evaluate, develop and deploy divers edge AI applications. For evaluation, a benchmark utility in Edge AI Suite can provide flexible settings to understand edge inference performance easily. Regarding development stage, one is to provide no-code/low-code platforms for creating retraining models easily and quickly. The other one is to provide available inference engine SDKs for various edge AI devices. With Custom Vision in Azure Cognitive Services, it’s easy to retrain an AI model for x86 platform to execute complex recognition. With Edge Impulse, it’s also easy to retrain an AI model but also optimized for ARM platform to execute specific recognition. To improve development efficiency of edge AI applications, ADVANTECH provides necessary and validated inference SDKs, like Intel OpenVINO and Nvidia TensorRT, for various edge AI hardware. Once customers’ edge AI applications are ready, with DeviceOn, we can remotely deploy or update AI applications to edge devices and monitor real-time status for them as well. Using DeviceOn Container Management, it’s simple to remotely monitor and operate containerized AI applications running at edge. To this end, ADVANTECH can provide not only one-stop-shop AI development kits to customers, but also cloud-to-edge AI architecture consulting service. With ADVANTECH hardware and Edge AI Suite, we helped reduce over fifty percent time spent for a factory production line to adopt an edge AI solution. In this case, the time for developing edge AI was reduced, edge AI deployment in the filed was accelerated, and edge AI management and maintenance was simplified. We also have lots of success stories for helping customers speed up edge AI application development and deployment. Just scan the QR code to submit your contact information. You will have AI solution consulting services for free from ADVANTECH.