A Distributed Principal Component Analysis Compression for Smart Seismic Acquisition Networks

Liu, Bo; Mohandes, M.; Nuha, Hilal Hudan; Deriche, Mohamed; Fekri, Faramarz

doi:10.1109/tgrs.2018.2789354

Cited by 30 publications

(6 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They all rely on the observation that the covariance matrix can be computed exactly at the global server by adding up the local covariance matrices. This basic version is, for instance, used in Liu et al (2018) . In this version, the covariance matrices of the local datasets are computed and sent to the aggregator.…”

Section: Methodsmentioning

confidence: 99%

Federated horizontally partitioned principal component analysis for biomedical applications

Hartebrodt

Röttger

2022

Bioinformatics Advances

View full text Add to dashboard Cite

Motivation Federated learning enables privacy preserving machine learning in the medical domain because the sensitive patient data remains with the owner and only parameters are exchanged between the data holders. The federated scenario introduces specific challenges related to the decentralized nature of the data, such as batch effects and differences in study population between the sites. Here, we investigate the challenges of moving classical analysis methods to the federated domain, specifically principal component analysis, a versatile and widely used tool, often serving as an initial step in machine learning and visualization workflows. We provide implementations of different federated PCA algorithms and evaluate them regarding their accuracy for high dimensional biological data using realistic sample distributions over multiple data sites, and their ability to preserve downstream analyses. Results Federated subspace iteration converges to the centralized solution even for unfavorable data distributions, while approximate methods introduce error. Larger sample sizes at the study sites lead to better accuracy of the approximate methods. Approximate methods may be sufficient for coarse data visualization, but are vulnerable to outliers and batch effects. Before the analysis, the PCA algorithm, as well as the number of eigenvectors should be considered carefully to avoid unnecessary communication overhead. Availability Simulation code and notebooks for federated PCA can be found at https://gitlab.com/roettgerlab/federatedPCA; the code for the federated app is available at https://github.com/AnneHartebrodt/fc-federated-pca Supplementary information Supplementary data are available at Bioinformatics Advances online.

show abstract

Section: Methodsmentioning

confidence: 99%

Federated horizontally partitioned principal component analysis for biomedical applications

Hartebrodt

Röttger

2022

Bioinformatics Advances

View full text Add to dashboard Cite

show abstract

“…where q is the smallest number of the eigenmodes we need to compute, i l is the lth largest singular value ofX i , and l is the lth largest singular value ofX. Equation (12) says that the singular values of the submatrices decay not only faster than that of the global matrix, but actually faster by a factor that is strictly less than 1. q does not appear explicitly in (12) but it is the lower bound for l. In other words, q is the smallest number of eigenmodes such that (12) holds.…”

Section: Some Analysis Of the Ddpcamentioning

confidence: 99%

Summation pollution of principal component analysis and an improved algorithm for location sensitive data

Cai

2021

Numerical Linear Algebra App

View full text Add to dashboard Cite

Principal component analysis (PCA) is widely used for dimensionality reduction and unsupervised learning. The reconstruction error is sometimes large even when a large number of eigenmode is used. In this paper, we show that this unexpected error source is the pollution effect of a summation operation in the objective function of the PCA algorithm. The summation operator brings together unrelated parts of the data into the same optimization and the result is the reduction of the accuracy of the overall algorithm. We introduce a domain decomposed PCA that improves the accuracy, and surprisingly also increases the parallelism of the algorithm. To demonstrate the accuracy and parallel efficiency of the proposed algorithm, we consider three applications including a face recognition problem, a brain tumor detection problem using two-and three-dimensional MRI images.

show abstract

“…Therefore, the extended network layer is responsible for strong coverage capability. The core network layer can be designed using cellular network technology (3G or 4G) and wired Ethernet technology [27]. However, seismic exploration usually works in complex areas where there are no 3G or 4G signals, which limits the use of cellular network technology.…”

Section: B Multi-hop Networking Solutionmentioning

confidence: 99%

Wireless Multi-Hop Energy-Efficient System for High-Density Seismic Array

et al. 2020

View full text Add to dashboard Cite

High-density seismic array method has attracted enormous interest in geophysical methods, and it is an advanced seismic observation technology which is of great significance in oil and gas resource detection and underground space exploration. However, designing a portable sensor node that prolongs the lifetime and equips wireless monitoring system to improve construction efficiency has been still a challenge for the high-density seismic array method. In this paper, a portable energy-efficient wireless sensor node and a hybrid communication system that content themselves with complex terrain, large-scale and reliable transmission are proposed to tackle the restrictions of high-density seismic survey. A Wi-Fi communication system based on the star network structure at the core network layer and a ZigBee communication system based on wireless multi-hop structure in the extended network layer form the hybrid communication system. In order to balance the energy of nodes in the extended network and prolong the network lifetime, this paper proposes a multi-hop variable weight routing method (MVWRM). The method selects the cluster-head node and the multi-hop route by changing the weight of the residual energy of the node and the weight of the distance at different periods. Then, a series of simulations and experiments are carried out for verification. The experiments show that the proposed sensor node suggest good performance in equivalent noise level (0.8µV@500Hz, PGA = 1), total harmonic distortion (124dB). Simulation results and performance analysis validate that the effectiveness of the proposed MVWRM can extend about 10% of the node lifetime. INDEX TERMS Energy-efficient, wireless sensor networks, multi-hop, high-density seismic array.

show abstract

A Distributed Principal Component Analysis Compression for Smart Seismic Acquisition Networks

Cited by 30 publications

References 19 publications

Federated horizontally partitioned principal component analysis for biomedical applications

Federated horizontally partitioned principal component analysis for biomedical applications

Summation pollution of principal component analysis and an improved algorithm for location sensitive data

Wireless Multi-Hop Energy-Efficient System for High-Density Seismic Array

Contact Info

Product

Resources

About