A physically-informed deep-learning model using time-reversal for locating a source from sparse and highly noisy sensors data

Kahana, Adar; Turkel, Eli; Dekel, Shai; Givoli, Dan

doi:10.1016/j.jcp.2022.111592

Cited by 7 publications

(3 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The recent developments of Transformer-based architectures have recently been used for this task as well (Ovadia et al, 2023 ). Another proposed method refers to using the Time-Reversal method incorporated with Machine learning based inference system (Bardos and Fink, 2002 ; Givoli and Turkel, 2012 ; Kahana et al, 2022 ). Most methods still rely on ANNs that can be expensive to train.…”

Section: Methodsmentioning

confidence: 99%

“…We venture into the domain of time-reversal wave localization problems, a significant challenge in physics and engineering. This problem aims to trace back a wave's source given the wave shape at a later time (Bardos and Fink, 2002 ; Givoli and Turkel, 2012 ; Kahana et al, 2022 ). Through these experiments, we aim to demonstrate the versatility and potential of SNNs in various complex tasks, significantly expanding their applicability beyond traditional domains.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Rethinking skip connections in Spiking Neural Networks with Time-To-First-Spike coding

Kim,

Kahana,

Yin

et al. 2024

Front. Neurosci.

Self Cite

View full text Add to dashboard Cite

Time-To-First-Spike (TTFS) coding in Spiking Neural Networks (SNNs) offers significant advantages in terms of energy efficiency, closely mimicking the behavior of biological neurons. In this work, we delve into the role of skip connections, a widely used concept in Artificial Neural Networks (ANNs), within the domain of SNNs with TTFS coding. Our focus is on two distinct types of skip connection architectures: (1) addition-based skip connections, and (2) concatenation-based skip connections. We find that addition-based skip connections introduce an additional delay in terms of spike timing. On the other hand, concatenation-based skip connections circumvent this delay but produce time gaps between after-convolution and skip connection paths, thereby restricting the effective mixing of information from these two paths. To mitigate these issues, we propose a novel approach involving a learnable delay for skip connections in the concatenation-based skip connection architecture. This approach successfully bridges the time gap between the convolutional and skip branches, facilitating improved information mixing. We conduct experiments on public datasets including MNIST and Fashion-MNIST, illustrating the advantage of the skip connection in TTFS coding architectures. Additionally, we demonstrate the applicability of TTFS coding on beyond image recognition tasks and extend it to scientific machine-learning tasks, broadening the potential uses of SNNs.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Rethinking skip connections in Spiking Neural Networks with Time-To-First-Spike coding

Kim,

Kahana,

Yin

et al. 2024

Front. Neurosci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The platform for different operations requires high accuracy of the navigation sensor [ 58 ]. According to the characteristics of DVL, the filtering technology based on sparse representation is adopted to reduce noise interference [ 59 ] and improve data accuracy.…”

Section: Control System Designmentioning

confidence: 99%

Design and Verification of Deep Submergence Rescue Vehicle Motion Control System

Jiang,

Zhang,

Wan

et al. 2023

Sensors

View full text Add to dashboard Cite

A six degree-of-freedom (DOF) motion control system for docking with a deep submergence rescue vehicle (DSRV) test platform was the focus of this study. The existing control methods can meet the general requirements of underwater operations, but the complex structures or multiple parameters of some methods have prevented them from widespread use. The majority of the existing methods assume the heeling effect to be negligible and ignore it, achieving motion control in only four or five DOFs. In view of the demanding requirements regarding positions and inclinations in six DOFs during the docking process, the software and hardware architectures of the DSRV platform were constructed, and then sparse filtering technology was introduced for data smoothing. Based on the adaptive control strategy and with a consideration of residual static loads, an improved S-plane control method was developed. By converting the force (moment) calculated by the controller to the body coordinate system, the complexity of thrust allocation was effectively reduced, and the challenge of thrust allocation in the case of a high inclination during dynamic positioning was solved accordingly. The automatic control of the trimming angle and heeling angle was realized with the linkage system of the ballast tank and pump valve. A PID method based on an intelligent integral was proposed, which not only dealt with the integral “saturation” problem, but also reduced the steady-state error and overshooting. Water pool experiments and sea trials were carried out in the presence of water currents for six-DOF motion control. The responsiveness and precision of the control system were verified by the pool experiment and sea trial results and could meet the control requirements in engineering practice. The reliability and operational stability of the proposed control system were also verified in a long-distance cruise.

show abstract

A comparison of machine learning methods for recovering noisy and missing 4D flow MRI data

Csala,

Amili,

D'Souza

et al. 2024

Numer Methods Biomed Eng

View full text Add to dashboard Cite

Experimental blood flow measurement techniques are invaluable for a better understanding of cardiovascular disease formation, progression, and treatment. One of the emerging methods is time‐resolved three‐dimensional phase‐contrast magnetic resonance imaging (4D flow MRI), which enables noninvasive time‐dependent velocity measurements within large vessels. However, several limitations hinder the usability of 4D flow MRI and other experimental methods for quantitative hemodynamics analysis. These mainly include measurement noise, corrupt or missing data, low spatiotemporal resolution, and other artifacts. Traditional filtering is routinely applied for denoising experimental blood flow data without any detailed discussion on why it is preferred over other methods. In this study, filtering is compared to different singular value decomposition (SVD)‐based machine learning and autoencoder‐type deep learning methods for denoising and filling in missing data (imputation). An artificially corrupted and voxelized computational fluid dynamics (CFD) simulation as well as in vitro 4D flow MRI data are used to test the methods. SVD‐based algorithms achieve excellent results for the idealized case but severely struggle when applied to in vitro data. The autoencoders are shown to be versatile and applicable to all investigated cases. For denoising, the in vitro 4D flow MRI data, the denoising autoencoder (DAE), and the Noise2Noise (N2N) autoencoder produced better reconstructions than filtering both qualitatively and quantitatively. Deep learning methods such as N2N can result in noise‐free velocity fields even though they did not use clean data during training. This work presents one of the first comprehensive assessments and comparisons of various classical and modern machine‐learning methods for enhancing corrupt cardiovascular flow data in diseased arteries for both synthetic and experimental test cases.

show abstract

A physically-informed deep-learning model using time-reversal for locating a source from sparse and highly noisy sensors data

Cited by 7 publications

References 25 publications

Rethinking skip connections in Spiking Neural Networks with Time-To-First-Spike coding

Rethinking skip connections in Spiking Neural Networks with Time-To-First-Spike coding

Design and Verification of Deep Submergence Rescue Vehicle Motion Control System

A comparison of machine learning methods for recovering noisy and missing 4D flow MRI data

Contact Info

Product

Resources

About