Lucas Talandier (CSIC)

The Post-Digital project significantly enriched Lucas’ PhD journey. Through a series of meetings, training sessions, and collaborations, it helped him build invaluable relationships that not only led to scientific breakthroughs but also profoundly deepened my understanding of unconventional computing, particularly in the realms of photonics and neuromorphic computing. Part of Lucas’ research involved numerical simulation, specifically working on simulating a photonic time-delay reservoir computer. Among the standout findings was the notable enhancement in signal recovery following nonlinear distortion in optical fibers, achieved by utilizing the phase information of received signals. These results were shared in an oral presentation at the SPIE Photonics West Conference in San Francisco by the supervising scientist in February 2023. Lucas’ numerical research continued by focusing on the time-delay reservoir computer with a shortened cavity, leading to two journal publications currently in preparation. On the experimental front, Lucas concentrated on characterizing nonlinearities in single-mode optical telecommunication fibers to improve signal recovery after nonlinear distortions using a reservoir computing device.

In collaboration with Irene Estebanez, a fellow PhD student from the IFISC nonlinear photonics group, he undertook experimental studies assessing the impact of phase versus intensity encoding on the reservoir’s input. These findings were also presented at the CLEO Europe 2023 conference in Munich by the supervising scientist.

Working with Tigers Jonuzi, a PhD student from VLC Photonics S.L., Lucas engaged in the characterization of a convolutional accelerator within a silicon photonic integrated chip at the Nonlinear Photonics Lab at IFISC. These findings have been accepted for presentation by Mr. Jonuzi at the upcoming SPIE Photonics West Conference in February 2024.

Lucas’ secondment at IBM Zurich involved working in collaboration mainly with Elger Vlieg on the project, to develop an integrated photonic chip that realizes multiply-accumulate (MAC) operations all-optically by using the photo-refraction effect. Lucas could through optimization of a free-space optics measurement setup, demonstrating two-wave mixing in the processed chip, culminating in a journal publication in collaboration with the supervising team at IBM.

Lucas’ secondment at VLC Photonics included extensive training in integrated photonic design, where he utilized leading industry tools to successfully simulate and design a convolutional network chip, which was subsequently experimentally characterized. Additionally, he played a significant role in characterizing an InP integrated photonics chip.