Professor Sergey Savel’ev of Loughborough University recently co-organised a special session on “Superconducting and high-frequency neuromorphic devices” at the 10th International Conference on Antennas and Electromagnetic Systems (AES 2024) held in Rome, Italy.
Superconducting and high-frequency neuromorphic devices can leverage quantum mechanical properties and operate at high speeds, making them ideal for implementing quantum reservoir computing such as those being developed in the QRC-4-ESP project.
During the session, Sergey Savel'ev and his Loughborough University colleagues gave a couple of invited talks: “Superconducting systems for quantum and neuromorphic AI and reservoir computing” from Sergey Savel'ev and “Artificial Neuron Circuit with a Two-Level Quantum Memristor” from Professor Alexander Balanov. Additionally, a poster was presented on “Bifurcation Phenomena in a Circuit with a Three-Level Quantum Memristor” from Finlay Potter, Alexandre Zagoskin, Sergey Savel'ev and Alexander Balanov.
In the first talk, they discussed alternative approaches for using time-series data in the training of potential AI superconducting layered systems, and also scrutinised diverse strategies for inputting data into the mesoscopic devices of layered superconductors.
In the second talk and the poster, they proposed a physical principle for the design of a quantum memristor, which could be used in quantum neuromorphic systems including reservoirs, and predicted the dynamics which artificial neurons based on such memristors could demonstrate.
Comentarios