Embedded AI@Edge – a Primer

Internet-of-Things, Robotics and AR/VR based applications are quite prevalent now – all of them employ constrained Embedded Edge Devices. With AI based analytics of the sensed data becoming an important part of Intelligent Sensing systems, designing and running AI on constrained embedded edge devices is becoming more and more important. In this talk we will introduce the need for AI-driven device edge computing, provide a few industrial use cases for it and finally discuss about the associated tools and technologies.

Speaker

Designing Deep Learning models for Embedded Systems

Different tools and frameworks are coming up for transforming/building AI/ML/DL models for embedded systems. Along with these, it is also very important to build models that adhere to the intended tasks and at the same time, are small in size, and fast while doing inference. Deep learning models are super complex, and a random architecture with unplanned training rarely generates a good model. One option is to reduce the models that were not built for embedded systems without changing their functionality. Another path is to understand how to design smaller models right from scratch. In this talk, we discuss these issues in depth.

Speaker

Tiny Edge– a tutorial

Resource-constrained devices such as microcontrollers are a mainstay in embedded systems. Recently, we see a great interest in the research community and industry to use these devices to do AI/ML inferencing tasks leading to embedded intelligence at the very edge. This is challenging and needs a deep knowledge of applications, algorithms, and computer architecture. In this hands-on session, we perform a deployment of a tiny model onto an edge system and explore options to enhance the performance using different means available.

SPEAKER