Explore an open-source standalone AI framework designed for almost any hardware in this tinyML Talks Germany presentation. Dive into AIfES (Artificial Intelligence for Embedded Systems), developed by Fraunhofer Institute for Microelectronic Circuits and Systems (IMS), which enables running and training artificial neural networks on a wide range of devices, including 8-bit microcontrollers. Learn about the framework's capabilities, its integration with Arduino IDE, and witness a live on-device training demonstration. Discover the latest features, licensing options, and potential applications of AIfES in embedded AI. Gain insights into the framework's architecture, data structures, and practical implementation through a gesture recognition demo, showcasing its versatility and efficiency in running machine learning on resource-constrained devices.
AIfES - An Open-Source Standalone AI Framework for Almost Any Hardware
Overview
Syllabus
Introduction
Small German Committee
Introducing Pierre
Embedded AI
Different solutions on the market
How does it work
Basic version
Hardware accelerators
Arduino IDE
Licensing
Paid license
Function overview
Feedforward Neural Network
Updates
Working with partners
Demonstration
Network description
Weight calculation
Training from scratch
Data structure
Gesture recognition
Demo
Confidence
Conclusion
Taught by
tinyML
