Artificial Intelligence (AI) is increasingly determining all of our lives and offers great potential for Germany's economic and innovative strength. However, the high energy consumption of AI technologies must be reduced. In order to promote new ideas for energy-saving chips, the Federal Ministry of Education and Research (BMBF) therefore initiated the pilot innovation competition “Energy-efficient AI systems”. For this reason, the German Federal Ministry of Education and Research (BMBF) initiated the pilot innovation competition "Energy-efficient AI systems" to promote new ideas for energy-saving chips. The task was to develop a chip that manages to detect cardiac arrhythmias and atrial fibrillation in ECG data with at least 90 percent accuracy while consuming the least energy.
The competition is one of three competitions initiated as a radar of ideas in the run-up to the establishment of the BMBF's Agency that supports Breakthrough Innovations. The agency is intended to be a flexible and rapid state funding instrument that supports and accelerates bringing the highly innovative ideas into the market.
In the ASIC 130 nanometer category (Application-Specific Integrated Circuit), the team from Friedrich-Alexander University Erlangen-Nuremberg and the team from the Fraunhofer Institute IIS Erlangen came out on top in the competition with the project: "Low-Power Low-Memory Low-Cost ECG Signal Analysis with ML Algorithms (Lo3-ML)".
In the project, the long-term partners of the Leibniz Institute for Innovative Microelectronics (IHP) developed a chip that uses non-volatile memories, so-called RRAMs - a technology developed by the IHP - as well as ultra-low-power circuits for writing and reading. This approach allows data to be recorded while the AI algorithm is inactive. To process the data, the algorithm is activated very quickly to complete its task in an extremely short time. In this way, the chip achieves energy savings of up to 95 percent compared to systems that are permanently active. The technology platform developed by IHP was one of three technology platforms used in the competition, along with the FPGA architecture solution and the process developed by GlobalFoundries.