Estimation of bubble size distribution using deep ensemble physics-informed neural network

Sunyoung Ko; Geunhwan Kim; Jaehyuk Lee; Hongju Gu; Kwangho Moon; Youngmin Choo

doi:10.7776/ASK.2023.42.4.305

All Issue

2023 Vol.42, Issue 4 Preview Page Next Page

Research Article

Estimation of bubble size distribution using deep ensemble physics-informed neural network 딥앙상블 물리 정보 신경망을 이용한 기포 크기 분포 추정

31 July 2023. pp. 305-312

PDF XML

Abstract

Physics-Informed Neural Network (PINN) is used to invert bubble size distributions from attenuation losses. By considering a linear system for the bubble population inversion, Adaptive Learned Iterative Shrinkage Thresholding Algorithm (Ada-LISTA), which has been solved linear systems in image processing, is used as a neural network architecture in PINN. Furthermore, a regularization based on the linear system is added to a loss function of PINN and it makes a PINN have better generalization by a solution satisfying the bubble physics. To evaluate an uncertainty of bubble estimation, deep ensemble is adopted. 20 Ada-LISTAs with different initial values are trained using the same training dataset. During test with attenuation losses different from those in the training dataset, the bubble size distribution and corresponding uncertainty are indicated by average and variance of 20 estimations, respectively. Deep ensemble Ada-LISTA demonstrate superior performance in inverting bubble size distributions than the conventional convex optimization solver of CVX.

Keywords

Deep ensemble

Physics-Informed Neural Network (PINN)

Adaptive Learned Iterative Shrinkage Thresholding Algorithm (Ada-LISTA)

Attenuation loss

Bubble size distribution

Uncertainty

기포 크기 분포를 음파 감쇄 손실을 이용하여 역산하기 위해 Physics-Informed Neural Network(PINN)을 사용하였다. 역산에 사용되는 선형시스템을 풀기 위해 이미지 처리 분야에서 선형시스템 문제를 해결한 Adaptive Learned Iterative Shrinkage Thresholding Algorithm(Ada-LISTA)를 PINN의 신경망 구조로 이용하였다. 더 나아가, PINN의 손실함수에 선형시스템 기반의 정규항을 포함함으로써 PINN의 해가 기포 물리 법칙을 만족하여 더 높은 일반화 성능을 가지도록 하였다. 그리고 기포 추정값의 불확실성을 계산하기 위해 딥앙상블 기법을 이용하였다. 서로 다른 초기값을 갖는 20개의 Ada-LISTA는 같은 훈련데이터를 이용하여 학습되었다. 이 후 테스트시 훈련데이터와 다른 경향의 감쇄 손실을 입력으로 사용하여 기포 크기 분포를 추정하였고, 추정값과 이에 대한 불확실성을 20개 추정값의 평균과 분산으로 각각 구하였다. 그 결과 딥앙상블이 적용된 Ada-LISTA는 기존 볼록 최적화 기법인 CVX보다 기포 크기 분포를 역산하는데 더 우수한 성능을 보였다.

키워드

딥앙상블

Physics-Informed Neural Network (PINN)

Adaptive Learned Iterative Shrinkage Thresholding Algorithm (Ada-LISTA)

감쇄 손실

기포 크기 분포

불확실성

References

H. Medwin and C. S. Clay, Fundamentals of Acoustical Oceanography (Academic Press, Cambridge, 1998), pp. 287-333. 10.1016/B978-012487570-8/50010-6

H. Medwin "Acoustical determination of bubble-size spectra," J. Acoust. Soc. Am. 62, 1041-1044 (1977). 10.1121/1.381617

R. Duraiswami, S. Prabhukumar, and G. L. Chahine, "Bubble counting using an inverse acoustic scattering method," J. Acoust. Soc. Am. 104, 2699-2717 (1998). 10.1121/1.423854

S. Boyd and L. Vandenberghe, Convex Optimization (Cambridge University Press, New York, 2004), pp. 16-23.

C. Park, S. W. Jeong, G. D. Kim, I. Moon, and G. Yim, "A study on the estimation of bubble size distribution using an acoustic inversion method" (in Korean), J. Acoust. Soc. Kr. 39, 151-162 (2020).

M. A. Ganaie, M. Hu, A. K. Malik, M. Tanveer, and P. N. Suganthan, "Ensemble deep learning: A review," Eng. Appl. Artif. lntell. 115, 105151 (2022). 10.1016/j.engappai.2022.105151

M. Raissi, P. Perdikaris, and G. E. Karniadakis, "Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations," J. Comput. Phys. 378, 686-707 (2019). 10.1016/j.jcp.2018.10.045

A. Aberdam, A. Golts, and M. Elad. "Ada-lista: Learned solvers adaptive to varying models," IEEE TPAMI/PAMI, 44, 9222-9235 (2021). 10.1109/TPAMI.2021.312504134735338

K. Gregor and Y. L. cun, Learning fast approximations of sparse coding. Proc. 27th International Conf. Machine Learning, 399-406 (2010).

K. Lee, C. Lee, and C. Park. "Insertion loss by bubble layer surrounding a spherical elastic shell submerged in water" (in Korean), J. Acoust. Soc. Kr. 41, 174-183 (2022).

Information

Publisher :The Acoustical Society of Korea
Publisher(Ko) :한국음향학회
Journal Title :The Journal of the Acoustical Society of Korea
Journal Title(Ko) :한국음향학회지
Volume : 42
No :4
Pages :305-312
Received Date : 2023-05-25
Accepted Date : 2023-07-04
DOI :https://doi.org/10.7776/ASK.2023.42.4.305

The Journal of the Acoustical Society of KoreaISSN:1225-4428(Print) 2287-3775(Online)한국음향학회

All Issue