PROMITHEAS: Photonic Reservoir cOMputing ImplemenTation aiming to Embed Artificial intelligence capabilitieS to next generation optical networks
Department of Informatics and Telecommunications, National and Kapodistrian University of Athens
In the context of the PROMITHEAS project the potential use of various micro/nano photonic structures for the realization of a full-scale all-optical reservoir-computing scheme will be investigated. Through advanced modeling tools the parameters that affect the classification performance of each photonic-neuron will be extracted, providing manufacturing-ready designs that will allow the realization of an optimized all-optical reservoir computer with increased performance, and reduced energy consumption compared to the conventional computer-based approaches. The optimized design will be employed in order to address sophisticated classification problems and allow the infiltration of artificial intelligence schemes into new areas like optical-networks. The applications that will be addressed consist of digital-word recognition in the optical domain without the use of opto-electronic conversions and with bit-rates of hundreds of Gbps, allowing the manufacturing of all-optical routers for backbone optical networks, and real-time video/image processing through the pure optical/analog processing capabilities of the proposed scheme.
Dimitris Syvridis, Professor, Department of Informatics and Telecommunications, University of Athens
Evangelos Grivas, Researcher, Department of Informatics and Telecommunications,University of Athens
Charis Mesaritakis, Researcher, Department of Informatics and Telecommunications,University of Athens