STMicroelectronics STM32Cube.AI Ecosystem Enhances Support for Efficient Machine Learning

     The STMicroelectronics STM32Cube.AI development environment provides users with a variety of machine learning techniques, giving them more flexibility to solve the three algorithmic challenges of classification, clustering and novelty detection as efficiently as possible.

     In addition to being able to develop neural networks for edge inference on STM32 microcontrollers (MCUs), the latest STM32Cube.AI version (version 7.0) supports new supervised and semi-supervised methods that can handle smaller datasets and fewer CPU cycles. These include Isolation Forest Anomaly Detection (iForest) and One-Class Support Vector Machine (OC-SVM) novelty detection, as well as K-means and SVM classifier algorithms, which users can now implement without human coding.

     In addition to neural networks, these classic machine learning algorithms allow developers to convert, validate and deploy various learning models on STM32 microcontrollers through easy-to-use techniques, shortening development cycles and solving AI development challenges faster.

     STM32Cube.AI allows developers to offload machine learning processing tasks from the cloud to STM32-based edge devices to reduce latency, save energy, improve cloud utilization, and protect privacy by minimizing data exchange over the Internet. Users now have additional flexibility with STM32 MCUs to choose efficient machine learning techniques for on-device analysis, ideal for long-term online use cases and smart battery-powered applications.