A conference paper developed within the DFG Collaborative Research Center 876 ("Data Analysis under Resource Constraints") by communication experts (Benjamin Sliwa, Christian Wietfeld from the Chair of Communication Networks of the Faculty ETIT) with experts in machine learning (Nico Piatkowski, former SFB 876 now ML2R) was awarded a Best Paper Award at the IEEE Flagship Conference "International Communications Conference (ICC)".

At the ICC 2020, which was originally planned for this year in Dublin, over 2100 papers were presented in a virtualized format. The SFB 876 paper awarded at the conference, entitled "LIMITS: Lightweight Machine Learning for IoT Systems with Resource Limitations", presents the novel open source framework LIghtweight Machine Learning for IoT Systems (LIMITS), which uses a platform-in-the-loop approach that explicitly takes into account the concrete software generation tools (the so-called compilation toolchain) of the Internet-of-Things (IoT) target platform.

LIMITS focuses on comprehensive tasks such as the automation of experiments and data acquisition, platform-specific code generation and the so-called sweet spot determination for optimal parameter combinations. In two case studies focusing on cellular data rate prediction and radio-based vehicle classification, LIMITS will be validated by comparing different learning models and real IoT platforms with memory constraints from 16 kB to 4 MB. Furthermore, its potential as a catalyst for the development of IoT systems with machine learning will be demonstrated.

https://www.kn.e-technik.tu-dortmund.de/cms/en/institute/News/2020/Best-Paper-ICC/index.html