AI tool automates machine learning process

Analog Devices Inc. (ADI) has recently unveiled a groundbreaking tool called AutoML for Embedded, designed to revolutionize edge AI development by streamlining the machine learning process from start to finish. Developed in collaboration with Antmicro, this AI tool is now an integral part of the Kenning framework, seamlessly integrated into CodeFusion Studio. The Kenning framework serves as a versatile and open-source platform dedicated to optimizing, benchmarking, and deploying AI models on edge devices.

By automating the end-to-end machine learning pipeline, AutoML for Embedded empowers developers of all backgrounds to create high-quality and efficient models without requiring extensive data science expertise. In a recent demonstration, this innovative tool was utilized to construct an anomaly detection model for sensory time series data on the ADI MAX32690 MCU. The AI model was successfully deployed on physical hardware as well as its digital twin in Renode simulation, showcasing its seamless integration and real-time performance monitoring capabilities.

As the demand for intelligent edge devices continues to rise and AI technology rapidly advances towards the edge, developers are faced with the challenge of fitting complex models onto compact microcontrollers. The complexities of data preprocessing, model selection, hyperparameter tuning, and hardware-specific optimizations present a steep learning curve for many developers. Recognizing this challenge, ADI collaborated on the development of AutoML for Embedded to empower developers to effortlessly build and deploy resource-intensive machine learning models on edge devices, such as microcontrollers, without the burden of complex coding or hardware constraints.

AutoML for Embedded, a Visual Studio Code plugin built on the Kenning library, offers support for a range of features including deployment to ADI MAX78002 AI accelerator MCUs and MAX32690 devices, simulation and RTOS workflows using Renode and Zephyr RTOS, and access to general-purpose open-source tools for flexible model optimization. The tool automates model search and optimization through cutting-edge algorithms, leveraging SMAC and Hyperband with Successive Halving to efficiently explore model architectures and training parameters, focusing resources on the most promising models while ensuring successful deployment by verifying model size against device RAM.

Michael Gielda, VP Business Development at Antmicro, emphasized the significance of this development, stating, "“Building on the flexibility of our open-source AI benchmarking and deployment framework, Kenning, we were able to develop an automated flow and VS code plugin that vastly reduces the complexity of building optimized edge AI models. Enabling workflows based on proven open-source solutions is the backbone of our end-to-end development services that help customers take full control of their product. With flexible simulation using Renode and seamless integration with the highly configurable and standardized Zephyr RTOS, the road to transparent and efficient edge AI development using AutoML in Kenning is open.”