Fraunhofer IMS, Arduino Release the AIfES Artificial Intelligence Framework for Almost All Arduinos

Fraunhofer IMS has announced a partnership with Arduino to release a high-efficiency framework for artificial intelligence, accessible directly within the Arduino Library Manager and compatible with boards all the way down to the resource-constrained Arduino Uno.

Building on the earlier AIfES (Artificial Intelligence for Embedded Systems) project, AIfES for Arduino delivers exactly what it sounds like: A variant of the AIfES framework for feed-forward neural networks (FFNs) compatible with the Arduino family of microcontrollers. As a result, Fraunhofer IMS boasts, AIfES now scales "from eight-bit microcontrollers over Raspberry Pi to smartphones or PCs."

"AIfES started in 2016 as a small project of mine to explore self-learning sensors," explains inventor Pierre Gembaczka, doctor of engineering. "I never thought that I would be able to introduce a whole framework together with Arduino.

"The Arduino Uno was the first board I tested AIfES on, so it's just a perfect fit. The computing power of the new Arduino Pro series will bring the possibilities to a whole new level. Being a maker myself, I’m incredibly excited to be able to share AIfES with the community to create exciting DIY projects."

"I believe this is a significant step for Machine Learning on MCUs: finally a framework designed from the ground up for constrained devices that delivers maximum performance while optimising power consumption," adds Arduino co-founder Massimo Banzi. "I can't wait to see what people will develop with this."

A key feature of the AIfES framework is that it doesn't require external connectivity - important, given the number of Arduino devices without network capabilities as standard. As well as being able to process sensor data through the network on-device, the framework allows for on-device training.

Fraunhofer IMS is making the AIfES for Arduino framework available under a dual-licensing model. Most will use the open source version, licensed under the reciprocal GNU General Public License 3; for those hoping to commercialize their creations without needing the result to also be made available under the GPL, a separate commercial license agreement is available for an unspecified sum.

The company has confirmed it plans to continue developing AIfES, both for Arduino and elsewhere, with a view to increasing its feature set. A future release will add support for convolutional neural networks (CNNs) alongside the current FNN support, and extend the data type support from 32-bit float (F32) to 32-bit integer (Q31) and eight-bit integer (Q7).

The framework is now available via the Arduino Library Manager in the Arduino IDE, with full source code published to GitHub under the GPL 3 license.