FAIRe aims to develop resource-limited AI for embedded systems, cyber-physical systems, and edge devices. The development approach should extend comprehensively from the application layer to the physical layer. The project will therefore investigate aspects of quantization, network optimization and precision adaptation. The system performance is measured by the ability of all abstraction layers to work together efficiently and to learn continuously and will be demonstrated by means of a human-machine interaction.
Partners
INRIA, DFKI
