INFOLANET INFORMATION PROCESSING WITH COUPLED LASER NETWORKS

In the INFOLANET project, we will combine the expertise of the PIs on dynamical systems and machine learning to advance information processing concepts, based on a high-speed photonic implementation. We anticipate that synchronization, understood in a broad sense, will be key to boost information processing performance. We aim to establish a network of coupled lasers that is able to generate diverse responses to the input stimulus (high-dimensionality of the responses) while keeping these responses reproducible for supervised learning (consistency of the responses). In order to comply with these apparently contradictory requirements, different aspects related to identical or generalized synchronization, stimulus-induced or autonomous synchronization, and transient or stable synchronization properties will need to be investigated to ultimately understand their impact on the performance.