Qeexo AutoML Shrinks Automated Machine Learning Footprint to Fit Cortex-M0(+)

Automated ML platform offers support for Arm's tiniest cores — an industry first.

Ish Ot Jr.
16 days agoMachine Learning & AI

As the power of machine learning migrates from large GPUs to mobile phones on down to mid-range microcontrollers, tools like Edge Impulse Studio enable developers to manage the entire pipeline in easy-to-use software, then deploy to Cortex-M7, M4, or even M3 targets. Qeexo, continuing in their similar quest to automate end-to-end machine learning for edge devices, have announced support for Arm's tiniest, most energy-efficient cores: the Cortex-M0 and Cortex-M0+.

The powerful tools available to today's makers often somewhat abstract the underlying hardware, so it may not always be clear what core the board you're using contains. Even within Arduino's own "Nano" line, MCUs range from the 8-bit ATmega328 and ATMega4809 in the original Nano and Nano Every, all the way up to the 32-bit Cortex-M4 core found in the Nano 33 BLE and Nano 33 BLE Sense's nRF52840 — a common target for tinyML applications. Somewhere in the middle is the Arduino Nano 33 IoT — an inexpensive Wi-Fi solution powered by a Microchip (née Atmel) SAM D21 — which is based on the smaller 32-bit Cortex-M0+ — and thus despite the similar name, often finds itself excluded from the many ML applications that its larger BLE siblings enjoy.

Other popular M0(+)-based devices, which have to date been largely been excluded from the machine learning party, include the tiny Tomu (deep neural networks that it in a USB port?!) Teensy-LC, BBC micro:bit and the immense quantity of ATSAMD21-based Trinkets, FeatherWings, and Arduino-compatibles from Adafruit.

Qeexo AutoML bringing tinyML to M0 and M0+ cores means that even the smallest, most power-conservative devices, such as smartwatches and other wearables, can easily gain intelligence, through a wide range of supported machine learning algorithms. Included in that list at present are GBM, XGBoost, Random Forest, Logistic Regression, Decision Tree, SVM, CNN, RNN, CRNN, ANN, Local Outlier Factor, and Isolation Forest.

In closing, a visionary quote from Dominic Pajak of VP Business Development at Arduino, and former product manager of the Cortex-M0 at Arm:

“Arduino is on a mission to make machine learning simple enough for anyone to use. We’re excited to partner with Qeexo AutoML to accelerate professional embedded ML development by guiding users to the optimal algorithms for their application. Combined with Arduino Nano 33 IoT, users can quickly create smart IoT sensors that can perform analytics at the edge, minimize communication, and maximize battery life.”
Related articles
Sponsored articles
Related articles