Design of real and complex recurrent neural networks for sound source localisation
Design of real and complex recurrent neural networks for sound source localisation
This study explores a pruning technique based on the singular value decomposition (SVD) for both real- and complex-valued recurrent neural networks (RNNs). Building on prior work on multilayer perceptrons (MLPs), the method reduces hidden-layer dimensions during training to reduce computation with minimal performance loss. Challenges unique to RNNs are addressed and compared to those of MLPs. Pruned models are compared with original architectures and a benchmark pruning method in a sound source localisation task. Results show real- and complex-valued RNNs can retain strong performance while using substantially fewer computational resources.
Paul, Vlad S.
3f46807b-cda2-4818-ba70-71bfd26a1a92
Nelson, Philip A.
5c6f5cc9-ea52-4fe2-9edf-05d696b0c1a9
2 March 2026
Paul, Vlad S.
3f46807b-cda2-4818-ba70-71bfd26a1a92
Nelson, Philip A.
5c6f5cc9-ea52-4fe2-9edf-05d696b0c1a9
Paul, Vlad S. and Nelson, Philip A.
(2026)
Design of real and complex recurrent neural networks for sound source localisation.
JASA Express Letters, 6 (3), [034801].
(doi:10.1121/10.0042761).
Abstract
This study explores a pruning technique based on the singular value decomposition (SVD) for both real- and complex-valued recurrent neural networks (RNNs). Building on prior work on multilayer perceptrons (MLPs), the method reduces hidden-layer dimensions during training to reduce computation with minimal performance loss. Challenges unique to RNNs are addressed and compared to those of MLPs. Pruned models are compared with original architectures and a benchmark pruning method in a sound source localisation task. Results show real- and complex-valued RNNs can retain strong performance while using substantially fewer computational resources.
Text
journal.pcbi.1013989
- Version of Record
More information
Accepted/In Press date: 12 February 2026
Published date: 2 March 2026
Identifiers
Local EPrints ID: 510240
URI: http://eprints.soton.ac.uk/id/eprint/510240
ISSN: 2691-1191
PURE UUID: f60aae84-72fc-467d-85b3-06a28749f697
Catalogue record
Date deposited: 24 Mar 2026 17:34
Last modified: 25 Mar 2026 03:15
Export record
Altmetrics
Contributors
Author:
Vlad S. Paul
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics