Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition

F. Martinini; A. Enttsel; A. Marchioni; M. Mangia; F. Pareschi; R. Rovatti; G. Setti

doi:10.1109/BioCAS54905.2022.9948688

Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition

F. Martinini, A. Enttsel, A. Marchioni, M. Mangia, F. Pareschi, R. Rovatti, G. Setti

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Deep Neural Networks (DNN) have become popular and widespread because they combine computational power and flexibility, but they may present critical hyper-parameters that need to be tuned before the model can be trained. Recently, the use of trainable binary masks in the field of Magnetic Resonance Imaging (MRI) acquisition brought new state-of-the-art results, but with the disadvantage of introducing a bulky hyper-parameter, which tuning is usually time-consuming. We present a novel callback-based method that is applied during training and turns the tuning problem into a triviality, also bringing non-negligible performance improvements. We test our method on the fastMRI dataset.

Original language	English (US)
Title of host publication	BioCAS 2022 - IEEE Biomedical Circuits and Systems Conference: Intelligent Biomedical Systems for a Better Future, Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	665-669
Number of pages	5
ISBN (Print)	9781665469173
DOIs	https://doi.org/10.1109/BioCAS54905.2022.9948688
State	Published - Jan 1 2022
Externally published	Yes

Access to Document

10.1109/BioCAS54905.2022.9948688

Cite this

Martinini, F., Enttsel, A., Marchioni, A., Mangia, M., Pareschi, F., Rovatti, R., & Setti, G. (2022). Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition. In BioCAS 2022 - IEEE Biomedical Circuits and Systems Conference: Intelligent Biomedical Systems for a Better Future, Proceedings (pp. 665-669). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/BioCAS54905.2022.9948688

@inproceedings{f8598a72e52e407d973db33e3f3e1ac6,

title = "Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition",

abstract = "Deep Neural Networks (DNN) have become popular and widespread because they combine computational power and flexibility, but they may present critical hyper-parameters that need to be tuned before the model can be trained. Recently, the use of trainable binary masks in the field of Magnetic Resonance Imaging (MRI) acquisition brought new state-of-the-art results, but with the disadvantage of introducing a bulky hyper-parameter, which tuning is usually time-consuming. We present a novel callback-based method that is applied during training and turns the tuning problem into a triviality, also bringing non-negligible performance improvements. We test our method on the fastMRI dataset.",

author = "F. Martinini and A. Enttsel and A. Marchioni and M. Mangia and F. Pareschi and R. Rovatti and G. Setti",

note = "Generated from Scopus record by KAUST IRTS on 2023-02-15",

year = "2022",

month = jan,

day = "1",

doi = "10.1109/BioCAS54905.2022.9948688",

language = "English (US)",

isbn = "9781665469173",

pages = "665--669",

booktitle = "BioCAS 2022 - IEEE Biomedical Circuits and Systems Conference: Intelligent Biomedical Systems for a Better Future, Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

address = "United States",

}

Martinini, F, Enttsel, A, Marchioni, A, Mangia, M, Pareschi, F, Rovatti, R & Setti, G 2022, Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition. in BioCAS 2022 - IEEE Biomedical Circuits and Systems Conference: Intelligent Biomedical Systems for a Better Future, Proceedings. Institute of Electrical and Electronics Engineers Inc., pp. 665-669. https://doi.org/10.1109/BioCAS54905.2022.9948688

Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition. / Martinini, F.; Enttsel, A.; Marchioni, A. et al.
BioCAS 2022 - IEEE Biomedical Circuits and Systems Conference: Intelligent Biomedical Systems for a Better Future, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2022. p. 665-669.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition

AU - Martinini, F.

AU - Enttsel, A.

AU - Marchioni, A.

AU - Mangia, M.

AU - Pareschi, F.

AU - Rovatti, R.

AU - Setti, G.

N1 - Generated from Scopus record by KAUST IRTS on 2023-02-15

PY - 2022/1/1

Y1 - 2022/1/1

N2 - Deep Neural Networks (DNN) have become popular and widespread because they combine computational power and flexibility, but they may present critical hyper-parameters that need to be tuned before the model can be trained. Recently, the use of trainable binary masks in the field of Magnetic Resonance Imaging (MRI) acquisition brought new state-of-the-art results, but with the disadvantage of introducing a bulky hyper-parameter, which tuning is usually time-consuming. We present a novel callback-based method that is applied during training and turns the tuning problem into a triviality, also bringing non-negligible performance improvements. We test our method on the fastMRI dataset.

AB - Deep Neural Networks (DNN) have become popular and widespread because they combine computational power and flexibility, but they may present critical hyper-parameters that need to be tuned before the model can be trained. Recently, the use of trainable binary masks in the field of Magnetic Resonance Imaging (MRI) acquisition brought new state-of-the-art results, but with the disadvantage of introducing a bulky hyper-parameter, which tuning is usually time-consuming. We present a novel callback-based method that is applied during training and turns the tuning problem into a triviality, also bringing non-negligible performance improvements. We test our method on the fastMRI dataset.

UR - https://ieeexplore.ieee.org/document/9948688/

UR - http://www.scopus.com/inward/record.url?scp=85142936781&partnerID=8YFLogxK

U2 - 10.1109/BioCAS54905.2022.9948688

DO - 10.1109/BioCAS54905.2022.9948688

M3 - Conference contribution

SN - 9781665469173

SP - 665

EP - 669

BT - BioCAS 2022 - IEEE Biomedical Circuits and Systems Conference: Intelligent Biomedical Systems for a Better Future, Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Martinini F, Enttsel A, Marchioni A, Mangia M, Pareschi F, Rovatti R et al. Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition. In BioCAS 2022 - IEEE Biomedical Circuits and Systems Conference: Intelligent Biomedical Systems for a Better Future, Proceedings. Institute of Electrical and Electronics Engineers Inc. 2022. p. 665-669 doi: 10.1109/BioCAS54905.2022.9948688

Training Binary Layers by Self-Shrinking of Sigmoid Slope: Application to Fast MRI Acquisition

Abstract

Access to Document

Other files and links

Fingerprint

Cite this