Advancing 6G-RAN through multi-technology, multi-sensor fusion, multi-band and multi-static perception
Scientific-Disciplinary Group
09/IINF-03 - Telecommunications
Description
• Regarding the training of spiking neural networks (SNN), the candidate will have to derivea lightweight surrogate version for the variable update rules of exact ADMM-basedtraining;• Will test and validate surrogate, hybrid, and exact ADMM updates, extending the trainingto multi-batch as well as full datasets;• Will collaborate in the extension of the exact ADMM update rules to Convolutional SNNlayers;• Will integrate the Anderson acceleration method into both surrogate and hybrid versions, asa means to reduce the number of iterations required to reach convergence;• Will implement the optimizer, including the new surrogate rules, the CNN layers andacceleration techniques, using PyTorch in full compatibility with the snnTorch framework;• Will characterize the performance of multi-layer dense networks in terms of accuracy andsparsity (computation), integrating CNN-layers as well into the study.
Compensation
22,500
Job posting website
Number of positions
1
Funding body
Dipartimento di Ingegneria dell'Informazione - Università degli Studi di Padova
How to apply
View the original posting on the MUR website: Go to MUR website