Learning to Generate AI

Small embedded systems plus artificial intelligence (AI) - the topic of our time. Learn AIfES is aimed at all those who deal with microcontrollers and electronic control systems as well as the implementation of AI.

We would like to show you in two webinars how to use our open source AI software framework AIfES (Artificial Intelligence for Embedded Systems) to execute and even train artificial neural networks (KNN) practically on any hardware. Particular focus is on running AI on simple microcontrollers and small IoT devices, so-called "tinyML".

• With Learn AIfES you have the chance to store your data on the device and perform the processing without transmission delay, with a significantly lower energy consumption compared to a PC. You thus achieve increased reliability through a decentralized architecture and can create a personalizable AI that optimizes itself based on your needs.

Learn AIfES supports not only the German government's AI strategywith the development of know-how and the promotion of young talent, but at the same time several UN sustainability goals, such as high-quality education, climate protection and industry, innovation and infrastructure. Be part of shaping a better environment and society for us and future generations.

www.aifes.ai

Check out AIfES on GitHub: https://github.com/Fraunhofer-IMS/AIfES_for_Arduino

The project is aimed at all makers, pupils, students or industrial employees who want to integrate AI in their projects and on their microcontrollers and electronic control systems.

With your support, we would like to hostfree webinars showing you how to implement a KNN on your microcontroller with AIfES and even train there. AIfES is completely implemented in the programming language C and therefore works on almost any hardware.

We show you everything you need to implement your ideas, from data acquisition and model creation to deployment and testing. For the training we use the classic Python tools and at the same time show you how to train with AIfES directly on the microcontroller.

It is also our goal to involve you in the process. For example, you can contribute your own ideas about which additional functions should be added to AIfES or participate in the development of a cool demonstrator to rebuild.

AI and machine learning are the topics of the future and AIfES is the only framework »Made in Germany« so far. With your support you help the AIfES team to bring the free webinars and possible other cool formats and developments to the world.

True to the open source idea, your donation enables the results to be made available to everyone else.

The project supports the UN Sustainable Development Goals education, climate action and innovations and infrastructure. Now, of course, you're wondering what AIfES has to do with fighting climate change? Deep learning on high-performance computers can emit a lot of CO2. In 2019, the University of Massachusetts did a life cycle assessment for training large AI models. The specific example involved the model-building process for natural language processing (NLP). They found that this process can emit more than 283.9 tons of CO₂. That's nearly five times the lifetime emissions of an average American car (source: https://arxiv.org/abs/1906.02243).

Every euro donated will be doubled by the Fraunhofer Future Foundation - until the first funding goal of € 30,000 is reached. So if you donate €5, the Future Foundation will double the amount to €10 until the limit of €30,000 is reached. After that, the funding takes place without doubling the donation.

When the first funding goal is reached, we will work out two free webinars, in which everyone can participate who wants to learn something about "tinyML" and AI. There we will show you in a crash course with simple examples everything you need to know to realize your own AI project with AIfES.

Actively participate and decide: When the second funding goal of 60.000 € is reached, new AIfES algorithms will be implemented or a new demonstrator will be developed, which you can rebuild. Here you can participate in a survey to decide which goal should be implemented.