SKM 2021 – wissenschaftliches Programm
HL 13.22: Poster
Dienstag, 28. September 2021, 13:30–16:30, P
Decision-Making Processes by a Kuramoto Model with Hebbian Learning: Circuit Synthesis and Wave Digital Emulation — •Sebastian Jenderny, Dennis Michaelis, and Karlheinz Ochs — Ruhr University Bochum, Chair of Digital Communication Systems, Bochum, Germany
Decision-making processes are an interesting topic often studied in synchronizing oscillatory networks. Here, synchronization is, on an abstract level, related to learning. In this context, the Hebbian learning rule can be interpreted as the increasing and decreasing coupling strength between oscillators with a small and a large phase difference, respectively. This can for example be implemented by the Kuramoto model, being a simple and well-studied model for oscillatory networks. Our aim is to synthesize an electrical circuit of the Kuramoto model with Hebbian learning with which decision-making processes can be mimicked. For this purpose, we derive a memristor model accounting for the Hebbian learning rule. We then develop a corresponding wave digital model and utilize it to mimic the decision-making process associated with the observation of optical illusions.