AI at the Edge: Solving Real-World Problems with Embedded Machine Learning by Daniel Situnayake and Jenny Plunkett.
Jenny holds BS in Electrical Engineering in from the University of Texas at Austin. She is a senior developer relations engineer at Edge Impulse. Daniel is Director of ML at Edge Impulse and holds a BSc Computer Networks and Security from Birmingham City University. He is the former Developer Advocate at Google for TensorFlow Lite.
This book is a guide to exploring how machine learning is being implemented on both edge devices and embedded systems. A bit of caution however, both authors are at Edge Impluse and their company product is referenced.
While not surprising, it can be viewed as a marketing product rather than addressing market leading solutions. The target audience is engineering professionals, including product managers and technology leaders.
We continue to deploy Internet of Things (IoT) devices. Many readers will think thermostats like Google Nest and X. However this focus is on industrial, automation, healthcare, agriculture, and autonomous vehicle devices which brings a lot of real-time data and machine learning driven decision-making at the remote device location. It is a very fast paced environment.
Edge AI will continue emerging as a key business driver
Daniel and Jenny do reveal workflows for developing edge AI applications. This does include outlining the exploration, goal setting, bootstrapping, testing, iteration, deployment, and support. They also provide guidance on building datasets, addressing ethical considerations, and implementing responsible AI practices. Actually there is a good review of AI ethics in the book. Finally they address strategies to evaluate, deploy, and support edge AI applications.
Systems with embedded machine learning are becoming more critical for business today. This is indeed transforming the way computers interact with real world data. When correctly deployed IoT devices can actually make decisions with their sensor data. Yet just five years ago this data was not used due to costs, bandwidth, and the limitations of IoT device power. However today we have ultra-low power microcontrollers to embedded Linux devices. Amazing changes will drive edge AI devices as agents become more integrated.
Consider this book: HBR’s 10 Must Reads on AI, Analytics, and the New Machine Age by Harvard Business Review:
In conclusion, AI at the Edge is a key market for machine learning, and data driven decision making. Again the target audience is engineering professionals, product managers and technology marketplace leaders.