POST-DIGITAL project would like to invite you to the


Summer School “Novel Nonlinear Substrates for Neural Networks”

15-16 September 2021

Besancon, France

This hybrid event is free and is organised by Université de Franche-ComtéThales and Université libre de Bruxelles.

Please note that the event will take place in Besancon, on the premises of FEMTO-ST Institute:

15B Avenue des Montboucons
25030 Besancon, France


For registration please contact Adriana Svetozarova

The time in the agenda is Brussels time.

  Day 1
15 September 2021
Electronic and individual atom systems
Day 2
16 September 2021
Photonic systems
9:00-10:00 Dr Alice Mizrahi
Thales, Paris,
“Spin torque oscillators for neural network computing”.
Prof Masaya Notomi
NTT Basic Research Laboratories / Tokyo Institute of Technology,
“Optoelectronic Accelerators based on Integrated Nanophotonics”.
10:00-11:00 Prof Alexander Khajetoorians
Radboud University Nijmegen, The Netherlands.
“What can we ‘learn’ from atoms? Going beyond neuromorphics”.
Prof Natalia G. Berloff,
University of Cambridge, UK.
“Unconventional computing with liquid light”.
11:00-11:30 Coffee Break Coffee Break
11:30-12:30 Matteo Cucchi,
Technische Universitaet Dresden, Germany
“Dendritic organic electrochemical networks for implantable
bioelectronics and reservoir computing”
ESR speaker 4-6
12:30-14:00 Lunch Break Lunch Break
14:00-14:20 ESR speaker 1 Prof Bhavin Shastri,
Queens University, Canada.
“Silicon photonics for neuromorphic computing and artificial intelligence”.
14:20-14:40 ESR speaker 2
14:40-15:00 ESR speaker 3
15:00-17:00 Poster session Discussion



Dr Alice Mizrahi, Thales, France: Spin-torque oscillators for neuromorphic computing

Spintronic devices exhibit a wide range of functionalities that are promising for neuromorphic computing. In particular, spin torque oscillators can act as artificial neurons due to their non-linear dynamics. In this talk I will provide several paths to leverage the non-linear dynamics of spin torque oscillators for computing, as well as describe the challenges that need to be addressed. I will focus on a particular architecture pushed by our group, a neural network where the spin torque neurons and synapses are connected through radiofrequency signals.

Prof Masaya Notomi, NTT Basic Research Laboratories / Tokyo Institute of Technology,Japan: Optoelectronic Accelerators based on Integrated Nanophotonics

Optical computing has reappeared on the stage of research and is expected to break the limitation of CMOS processors, especially for analog computations. To fully utilize the merit of optics in computing, however, it is essentially important to combine optic and electronic circuits very efficiently. Here we show that integrated nanophotonics enables extremely-efficient optical-to-electrical and digital-to-analogue conversion. We also show how to implement them in optoelectronic accelerators enabling ultralow-latency computations.

Prof Alexander Khajetoorians, Radboud University Nijmegen, The Netherlands: What can we ‘learn’ from atoms? Going beyond neuromorphics

The quest to implement machine learning algorithms in hardware has focused on combining various materials, each mimicking a computational primitive, to create device functionality. These endeavors have led to the beautiful development of dedicated hardware that, working in combination with software, can perform pattern recognition tasks. Ultimately, these piecewise approaches limit functionality and efficiency, while complicating scaling and on-chip learning, necessitating new approaches linking physical phenomena to machine learning models. Likewise, this raises the question if there are new machine learning algorithms to be discovered, utilizing the particular properties of quantum properties of matter where there are no obvious links to established models. Here, I will discuss the first steps toward a new paradigm in computing, routed in fundamentals studies based on the idea of letting the physics do the work. I will introduce the concept of an atomic orbital memory and how coupling leads to tunable multi-modal landscapes. I will discuss how the ensuing stochastic dynamics mimics the Boltzmann machine, scaled to just seven atoms. In this discussion, I will review the emergence of multiple and separable time scales, an adaption of long-term potentiation in biological matter, which serves the basis for self-adaption and on-chip learning. I will conclude with an outlook on concepts that go beyond the current neuromorphic paradigm, combining concepts related to quantum coherent and quantum technologies.

Matteo Cucchi, Technical University Dresden, Germany: Dendritic organic electrochemical networks for implantable bioelectronics and reservoir computing

Early detection of malign patterns in patients’ biological signals can save millions of lives. Despite the steady improvement of AI-based techniques, the practical clinical application of these methods is mostly constrained to an offline evaluation of the patients’ data. Previous studies have identified organic electrochemical devices as ideal candidates for biosignal monitoring. However, their use for pattern recognition in real-time was never demonstrated. I will present the production, characterization, and use of brain-inspired networks composed of organic electrochemical transistors for time-series predictions and classification tasks using the reservoir computing approach.

Prof Natalia Berloff, University of Cambridge, UK: Unconventional computing with liquid light

The recent advances in developing physical platforms for solving combinatorial optimisation problems reveal the future of high-performance computing for quantum and classical devices. Unconventional computing architectures were proposed for numerous optical systems, including parametric oscillators, memristors, lasers and nanolasers, optoelectronic systems, photonic simulators, trapped ions, polariton and photon condensates. A promising approach to achieve computational supremacy over the classical von Neumann architecture explores classical and quantum hardware as Ising and XY machines. Gain-dissipative platforms such as the networks of optical parametric oscillators, coupled lasers and non-equilibrium Bose-Einstein condensates such as exciton-polariton or photon condensates use an approach to finding the global minimum of spin Hamiltonians which is different from quantum annealers or quantum computers. In my talk, I will discuss the principles of the operation of the devices based on such systems, the challenges they present, and the question of comparing different platforms’ performance.

Prof Bhavin Shastri, Queen’s University, Canada: Silicon photonics for neuromorphic computing and artificial intelligence

Artificial intelligence enabled by neural networks has enabled applications in many fields (e.g. medicine, finance, autonomous vehicles). Software implementations of neural networks on conventional computers are limited in speed and energy efficiency. Neuromorphic engineering aims to build processors in which hardware mimic neurons and synapses in brain for distributed and parallel processing. Neuromorphic engineering enabled by silicon photonics can offer sub nanosecond latencies, and can extend the domain of artificial intelligence and neuromorphic computing applications to machine learning acceleration (vector-matrix multiplications, inference and ultrafast training), nonlinear programming (nonlinear optimization problem and differential equation solving) and intelligent signal processing (wideband RF and fiber-optic communications). We will discuss current progress and challenges of neuromorphic photonics to scale to practical systems.