The latest release of Qeexo AutoML makes the automated TinyML model development and deployment platform available as a web application hosted on Amazon Web Services (AWS). Other significant features of the release include:

• Support for RNNs, Isolation Forests, and Local Outlier Factor algorithms
• A classifier optimized for anomaly detection
• Manual selection of sensors and sensor data features used in model building after data has been recorded
• The ability to visualizes classes separately
• Classification sensitivity tuning using visualization and sensitivity analysis tools
• Neural network parameter configurations (such as quantization-aware training)

Qeexo AutoML has currently been ported to the Arduino Nano 33 BLE Sense, Renesas RA6M3 ML Sensor Module, and STMicroelectronics STWINKT1 and SensorTile.box.

For a limited time, users can register for a free Bronze package of AutoML that allows them to upload or collect data sets and automatically build and deploy lightweight machine learning models onto embedded hardware targets.

For more information, visit http://automl.qeexo.com.

To see the Qeexo AutoML platform in action, watch the video below.