scholarly journals Solving inverse problems using data-driven models

Acta Numerica ◽  
2019 ◽  
Vol 28 ◽  
pp. 1-174 ◽  
Author(s):  
Simon Arridge ◽  
Peter Maass ◽  
Ozan Öktem ◽  
Carola-Bibiane Schönlieb
Keyword(s):  
Inverse Problems ◽  
Industrial Applications ◽  
Data Driven ◽  
Analytical Models ◽  
Survey Paper ◽  
Domain Specific ◽  
Combining Data ◽  
Domain Specific Knowledge ◽  
Ill Posed ◽  
Using Data

Recent research in inverse problems seeks to develop a mathematically coherent foundation for combining data-driven models, and in particular those based on deep learning, with domain-specific knowledge contained in physical–analytical models. The focus is on solving ill-posed inverse problems that are at the core of many challenging applications in the natural sciences, medicine and life sciences, as well as in engineering and industrial applications. This survey paper aims to give an account of some of the main contributions in data-driven inverse problems.

scholarly journals COMBINING DATA-DRIVEN AND NUMERICAL MODELLING APPROACHES TO STORM EROSION PREDICTION

2020 ◽  
pp. 38
Author(s):  
Joshua Simmons ◽  
Kristen Splinter
Keyword(s):  
Numerical Models ◽  
Individual Performance ◽  
Study Data ◽  
Data Driven ◽  
Erosion Prediction ◽  
Trade Offs ◽  
Combining Data ◽  
Erosion Models ◽  
Using Data ◽  
The Individual

Physics-based numerical models play an important role in the estimation of storm erosion, particularly at beaches for which there is little historical data. However, the increasing availability of pre-and post-storm data for multiple events and at a number of beaches around the world has opened the possibility of using data-driven approaches for erosion prediction. Both physics-based and purely data-driven approaches have inherent strengths and weaknesses in their ability to predict storm-induced erosion. It is vital that coastal managers and modelers are aware of these trade-offs as well as methods to maximise the value from each modelling approach in an increasingly data-rich environment. In this study, data from approximately 40 years of coastal monitoring at Narrabeen-Collaroy Beach (SE Australia)has been used to evaluate the individual performance of the numerical erosion models SBEACH and XBeach, and a data-driven modelling technique. The models are then combined using a simple weighting technique to provide a hybrid estimate of erosion.Recorded Presentation from the vICCE (YouTube Link): https://youtu.be/v53dZiO8Y60

scholarly journals Mammogram Segmentation using Multi-atlas Deformable Registration

2019 ◽  
Author(s):  
Manish Kumar Sharma ◽  
Mainak Jas ◽  
Vikrant Karale ◽  
Anup Sadhu ◽  
Sudipta Mukhopadhyay
Keyword(s):  
Ground Truth ◽  
Data Driven ◽  
Second Phase ◽  
Morphological Operations ◽  
Region Segmentation ◽  
Expert Radiologist ◽  
Candidate Image ◽  
Combining Data ◽  
Using Data ◽  
Automated Algorithms

Accurate breast region segmentation is an important step in various automated algorithms involving detection of lesions like masses and microcalcifications, and efficient telemammography. Existing segmentation algorithms underperform due to variations in image quality and shape of the breast region. In this paper, we propose to segment breast region by combining data-driven clustering with deformable image registration. In the first phase of the approach, we identify atlas images from a dataset of mammograms using data-driven clustering. Then, we segment these atlas images and use in the next phase of the algorithm. The second phase is atlas-based registration. For a candidate image, we find the most similar atlas image from the set of atlases identified in phase one. We deform the selected atlas image to match the given test image using the Demon's registration algorithm. Then, the segmentation mask of the deformed atlas is transferred to the mammogram in consideration. Finally, we refine the segmentation mask with some morphological operations in order to obtain accurate breast region boundary. We evaluated the performance of our method using ground-truth segmentation masks verified by an expert radiologist. We compared the proposed method with three existing state-of-the-art algorithms for breast region segmentation and the proposed approach outperformed all three in most of the cases.

Physical modelling of industrial multistage centrifugal compressors for monitoring and simulation

2015 ◽  
Vol 229 (18) ◽  
pp. 3433-3448 ◽  
Author(s):  
Matteo Cicciotti ◽  
Dionysios P Xenos ◽  
Ala EF Bouaswaig ◽  
Nina F Thornhill ◽  
Ricardo F Martinez-Botas
Keyword(s):  
Physical Modelling ◽  
Industrial Applications ◽  
Data Driven ◽  
Variable Speed ◽  
One Dimensional ◽  
Simulation And Optimization ◽  
The Common ◽  
Using Data ◽  
Multistage Centrifugal Compressor

Currently, industrial applications to monitoring, simulation and optimization of compressors employ empirical models that are either data-driven or based on the manufacturer performance maps. This paper proposes the use of one-dimensional aerodynamic models for industrial applications such as simulation and monitoring. The physical model establishes causality relationships among input and output variables that are tuned to match the real compressor by using operation data. The application of the method is shown using data from an industrial multistage centrifugal compressor with interstage coolers and variable inlet guide vanes. This is a more complex but more relevant case study for process industry, as opposed to the single-stage variable speed compressors, which is the common example in the literature.

scholarly journals Conditional Invertible Neural Networks for Medical Imaging

2021 ◽  
Vol 7 (11) ◽  
pp. 243
Author(s):  
Alexander Denker ◽  
Maximilian Schmidt ◽  
Johannes Leuschner ◽  
Peter Maass
Keyword(s):  
Neural Networks ◽  
Medical Imaging ◽  
Inverse Problems ◽  
Low Dose ◽  
Data Driven ◽  
Resonance Imaging ◽  
Point Estimates ◽  
Ill Posed ◽  
Low Dose Computed Tomography

Over recent years, deep learning methods have become an increasingly popular choice for solving tasks from the field of inverse problems. Many of these new data-driven methods have produced impressive results, although most only give point estimates for the reconstruction. However, especially in the analysis of ill-posed inverse problems, the study of uncertainties is essential. In our work, we apply generative flow-based models based on invertible neural networks to two challenging medical imaging tasks, i.e., low-dose computed tomography and accelerated medical resonance imaging. We test different architectures of invertible neural networks and provide extensive ablation studies. In most applications, a standard Gaussian is used as the base distribution for a flow-based model. Our results show that the choice of a radial distribution can improve the quality of reconstructions.

On a characteristic property of conditionally well-posed problems

2015 ◽  
Vol 23 (3) ◽  
Author(s):  
Mikhail Y. Kokurin
Keyword(s):  
Inverse Problems ◽  
Input Data ◽  
Characteristic Property ◽  
Optimization Problems ◽  
Data Driven ◽  
Error Models ◽  
Regularizing Algorithms ◽  
Stochastic Error ◽  
Ill Posed ◽  
Well Posed

AbstractThe aim of this paper is to discuss and illustrate the fact that conditionally well-posed problems stand out among all ill-posed problems as being regularizable via an operator independent of the level of errors in input data. We give examples of corresponding purely data driven regularizing algorithms for various classes of conditionally well-posed inverse problems and optimization problems in the context of deterministic and stochastic error models.

scholarly journals A New Hybrid Method for Solving Inverse Heat Conduction Problems

2021 ◽  
Vol 15 ◽  
pp. 151-158
Author(s):  
M. R. Shahnazari ◽  
F. Roohi Shali ◽  
A. Saberi ◽  
M. H. Moosavi
Keyword(s):  
Heat Conduction ◽  
Inverse Problems ◽  
Regularization Parameter ◽  
Accurate Method ◽  
Fundamental Solutions ◽  
Industrial Applications ◽  
Inverse Heat Conduction ◽  
Constant Coefficients ◽  
Ill Posed ◽  
Inverse Heat Conduction Problems

Solving the inverse problems, especially in the field of heat transfer, is one of the challenges of engineering due to its importance in industrial applications. It is well-known that inverse heat conduction problems (IHCPs) are severely ill-posed, which means that small disturbances in the input may cause extremely large errors in the solution. This paper introduces an accurate method for solving inverse problems by combining Tikhonov's regularization and the genetic algorithm. Finding the regularization parameter as the decisive parameter is modelled by this method, a few sample problems were solved to investigate the efficiency and accuracy of the proposed method. A linear sum of fundamental solutions with unknown constant coefficients assumed as an approximated solution to the sample IHCP problem and collocation method is used to minimize residues in the collocation points. In this contribution, we use Morozov's discrepancy principle and Quasi-Optimality criterion for defining the objective function, which must be minimized to yield the value of the optimum regularization parameter.

Development of a Pediatric Early Warning System Using Data-Driven Vital Signs

PEDIATRICS ◽  
2016 ◽  
Vol 137 (Supplement 3) ◽  
pp. 256A-256A
Author(s):  
Catherine Ross ◽  
Iliana Harrysson ◽  
Lynda Knight ◽  
Veena Goel ◽  
Sarah Poole ◽  
...  
Keyword(s):  
Early Warning ◽  
Early Warning System ◽  
Vital Signs ◽  
Warning System ◽  
Data Driven ◽  
Using Data

‘I’m an expert in me and I know what I can cope with’: Patient expertise in rheumatoid arthritis

2014 ◽  
Vol 10 (3) ◽  
pp. 249-261 ◽  
Author(s):  
Tessa Sanderson ◽  
Jo Angouri
Keyword(s):  
Rheumatoid Arthritis ◽  
Meta Analysis ◽  
Healthcare Systems ◽  
Specific Knowledge ◽  
Active Involvement ◽  
Domain Specific ◽  
Access To Health ◽  
Patient Expertise ◽  
Domain Specific Knowledge ◽  
The Uk

The active involvement of patients in decision-making and the focus on patient expertise in managing chronic illness constitutes a priority in many healthcare systems including the NHS in the UK. With easier access to health information, patients are almost expected to be (or present self) as an ‘expert patient’ (Ziebland 2004). This paper draws on the meta-analysis of interview data collected for identifying treatment outcomes important to patients with rheumatoid arthritis (RA). Taking a discourse approach to identity, the discussion focuses on the resources used in the negotiation and co-construction of expert identities, including domain-specific knowledge, access to institutional resources, and ability to self-manage. The analysis shows that expertise is both projected (institutionally sanctioned) and claimed by the patient (self-defined). We close the paper by highlighting the limitations of our pilot study and suggest avenues for further research.

Sign in / Sign up

Export Citation Format

Share Document