Corpus ID: 209862603

The troublesome kernel: why deep learning for inverse problems is typically unstable

  title={The troublesome kernel: why deep learning for inverse problems is typically unstable},
  author={Nina Gottschling and Vegard Antun and B. Adcock and A. Hansen},
  • Nina Gottschling, Vegard Antun, +1 author A. Hansen
  • Published 2020
  • Computer Science
  • ArXiv
  • There is overwhelming empirical evidence that Deep Learning (DL) leads to unstable methods in applications ranging from image classification and computer vision to voice recognition and automated diagnosis in medicine. Recently, a similar instability phenomenon has been discovered when DL is used to solve certain problems in computational science, namely, inverse problems in imaging. In this paper we present a comprehensive mathematical analysis explaining the many facets of the instability… CONTINUE READING
    17 Citations
    Interval Neural Networks as Instability Detectors for Image Reconstructions
    • Highly Influenced
    • PDF
    Deep Learning Techniques for Inverse Problems in Imaging
    • 46
    • PDF
    Model Adaptation for Inverse Problems in Imaging
    • PDF
    The gap between theory and practice in function approximation with deep neural networks
    • 7
    • PDF
    Wavelets in the Deep Learning Era
    • PDF
    Stabilizing Deep Tomographic Reconstruction Networks
    • 4
    • PDF
    Deep learning for biomedical image reconstruction: a survey
    • 2
    • PDF


    On instabilities of deep learning in image reconstruction and the potential costs of AI
    • 73
    • PDF
    Learning The Invisible: A Hybrid Deep Learning-Shearlet Framework for Limited Angle Computed Tomography
    • 49
    • PDF
    Intriguing properties of neural networks
    • 6,070
    • PDF
    Some Investigations on Robustness of Deep Learning in Limited Angle Tomography
    • 30
    • PDF
    Deep Convolutional Neural Network for Inverse Problems in Imaging
    • 852
    • Highly Influential
    • PDF
    Solving ill-posed inverse problems using iterative deep neural networks
    • 233
    • PDF
    DeepFool: A Simple and Accurate Method to Fool Deep Neural Networks
    • 2,048
    • PDF
    Convolutional Neural Networks for Inverse Problems in Imaging: A Review
    • 269
    • PDF
    Deep Learning: An Introduction for Applied Mathematicians
    • 70
    • PDF