Mini-symposium: Training of physical NN beyond back propagation (13 January, 2026) ULB, Brussels
Tue 13 Jan
|Brussels
A focused session on backpropagation-free training algorithms for physical neural networks, covering convergence, hardware requirements, substrate limitations, and potential applications.
Time & Location
13 Jan 2026, 09:00 – 18:00
Brussels
About the Event
POSTDIGITAL+ Mini-Symposium: Training of Physical Neural Networks Beyond Backpropagation
Date: 13 January 2026
Location: Université libre de Bruxelles (ULB), Brussels
On 13 January 2026, the POSTDIGITAL+ Doctoral Network organised a Mini-Symposium on “Training of Physical Neural Networks Beyond Backpropagation” at the Université libre de Bruxelles. The event brought together leading researchers from academia and industry to examine emerging approaches for training physical neural networks, with a particular emphasis on photonic and neuromorphic computing systems.
The symposium featured a series of invited talks addressing both theoretical and applied perspectives on next-generation machine learning hardware. Topics ranged from system-level simulation and architecture design to model-free learning strategies and alternative training paradigms.
Invited speakers and topics included: