Tigers Jonuzi (VLC)

In POST-DIGITAL, the work of Tigers Jonuzi focused on integrated photonics and neuromorphic network architectures, investigating their principles and applications. The focus of the methodology lies in the study and development of actuation mechanisms for optical weight assignation, including phase shifters, thermo-optic and graphene actuators, as well as the integration of non-volatile actuation with phase change materials. Additionally, the integration of Micro-electromechanical Systems (MEMS) into silicon photonics is examined.

His main research contributions explore wavelength and time space multiplexing architectures. The investigation of wavelength dimensionality includes topics such as wavelength multiplexing and the generation of frequency combs [1]. The integration of wavelength and time multiplexing allows for massive parallelization. The information processed in different wavelength is easier to control avoiding phase-dependent weights increasing the robustness of the architectures [2]. Furthermore, he investigates time-multiplexed reservoir computing and introduces a photonic convolutional accelerator [3], detailing its design and providing preliminary experimental setups and characterizations. The tests demonstrate an architecture which is limited only by the electronic hardware speed capable of supporting information processing at any data rates [4].  This innovation holds great promise in enhancing the computational efficiency of neuromorphic systems, bridging the gap between photonics and neural network processing.

Tigers work advances the understanding and application of Integrated Photonics and Neuromorphic Network Architectures, with significant contributions in exploring multiplexing architectures for enhanced computing and information processing systems.

Key publications by Tigers Jonuzi:

[1] Tigers Jonuzi, Alessandro Lupo, Miguel C. Soriano, J. David Domenech Gomez, and Serge Massar “Integrated optical output layer for a reservoir computer based on frequency multiplexing”, Proc. SPIE 12438, AI and Optical Data Sciences IV, 124380C (15 March 2023);

[2] Tigers Jonuzi, Alessandro Lupo, Miguel C. Soriano, Serge Massar, and J. D. Domenéch, “Integrated programmable spectral filter for frequency-multiplexed neuromorphic computers,” Opt. Express 31, 19255-19265 (2023);

[3] T. Jonuzi, S. Masaad, A. Lupo, J. D. Doménech Gomez, P. Bienstman and S. Massar, “Numerical Analysis of a Self-Calibrating Time-Spatial Interleaving Photonic Convolutional Accelerator,”2023 International Conference on Photonics in Switching and Computing (PSC), Mantova, Italy, 2023, pp. 1-4;

[4] Tigers Jonuzi, Alessandro Lupo, Lucas Talandier, Mirko Goldmann, Ingo Fischer, Apostolos Argyris, Serge Massar, Miguel C. Soriano, J. David Domenech Gomez, “Reconfigurable time-space photonic integrated convolutional accelerator,” Proc. SPIE 12903, AI and Optical Data Sciences V, 1290304 (13 March 2024)