Superconducting artificial synapses integrated into a self-training neuromorphic architecture

Emilie Jué (National Institute of Standards and Technology, Colorado, USA)

Date: May 14^th, 2025

Superconducting electronics is a compelling technology for designing neuromorphic computing systems. The technology uses Josephson junctions (JJ), which have a natural spiking behavior and can transmit voltage spikes over long distances with near-zero loss. One important building block of a neural network that still needs to be developed in superconducting electronics is the synapse, whose role is to adjust the strength of the connection between two neurons by changing the synaptic weight. In this work, we propose a new superconducting synapse developed at NIST and demonstrate its integrations in neural networks. The artificial synapse is obtained with a SQUID-based circuit using a flux storage loop for the synaptic weight. Using SPICE simulations, we demonstrate that the synaptic
circuit can be tuned using digital single flux quantum pulses.

In the second part of this presentation, we demonstrate the implementation of the synapses in a neural network. We integrate the SQUID-based artificial synapses into a small-scale self-training neural network architecture using SPICE simulations. The network follows reinforcement learning rules that update local weights internally. This property allows the network to learn new functions by changing the target output for a given set of inputs without needing any external adjustments. Finally, using the same constraint as for the network mentioned above, we extend our simulations in Python to a larger problem to classify the handwritten digits from the MNIST dataset and show that the MNIST network can be trained in about 100 ms.

Neuromorphic Computing

Frank Mizrahi (Laboratoire Albert Fert, CNRS, Thales, Université Paris-Saclay)

Date: February 26^th, 2025

The ability to compute and learn at the edge is critical for many artificial intelligence applications, from medical sensors to autonomous vehicles. Yet, this is not possible with current hardware because of the high energy consumption of learning. Neuromorphic computing aims at developing novel energy efficient hardware by taking inspiration from the brain. In this seminar, we will see what ideas from the brain can be used to develop novel hardware. We will explain the main challenges of the field and review the emerging technologies studied.

Magnetization dynamics features in φ0 Josephson junctions

Andrei Mazanik (CFM-CSIC)

Date: November 6^th, 2024

The consortium is proud to present a series of seminars called Josephine Seminars, in which consortium members, as well as external guests working on topics related to the project’s objectives, will give scientific talks. The seminars will be online and open, i.e. scientists from all over the world can participate. Each seminar will last 45 minutes plus 15 minutes for questions. You can attend the semianr via Zoom. Between 3 and 4 seminars will be organized annually. This is the first one of the series.