Deep learning methods for solving linear inverse problems: Research directions and paradigms
The linear inverse problem is fundamental to the development of various scientific areas.
Innumerable attempts have been carried out to solve different variants of the linear inverse …
Innumerable attempts have been carried out to solve different variants of the linear inverse …
Learning to optimize: A primer and a benchmark
Learning to optimize (L2O) is an emerging approach that leverages machine learning to
develop optimization methods, aiming at reducing the laborious iterations of hand …
develop optimization methods, aiming at reducing the laborious iterations of hand …
The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and Smale's 18th problem
MJ Colbrook, V Antun… - Proceedings of the …, 2022 - National Acad Sciences
Deep learning (DL) has had unprecedented success and is now entering scientific
computing with full force. However, current DL methods typically suffer from instability, even …
computing with full force. However, current DL methods typically suffer from instability, even …
Learning to optimize: A tutorial for continuous and mixed-integer optimization
Learning to optimize (L2O) stands at the intersection of traditional optimization and machine
learning, utilizing the capabilities of machine learning to enhance conventional optimization …
learning, utilizing the capabilities of machine learning to enhance conventional optimization …
Hyperparameter tuning is all you need for LISTA
Abstract Learned Iterative Shrinkage-Thresholding Algorithm (LISTA) introduces the concept
of unrolling an iterative algorithm and training it like a neural network. It has had great …
of unrolling an iterative algorithm and training it like a neural network. It has had great …
Adversarial robustness of supervised sparse coding
J Sulam, R Muthukumar… - Advances in neural …, 2020 - proceedings.neurips.cc
Several recent results provide theoretical insights into the phenomena of adversarial
examples. Existing results, however, are often limited due to a gap between the simplicity of …
examples. Existing results, however, are often limited due to a gap between the simplicity of …
Deep Parametric Imaging for Bistatic SAR: Model, Property and Approach
Y Song, W Pu, J Huo, J Wu, Z Li… - IEEE Transactions on …, 2024 - ieeexplore.ieee.org
Bistatic synthetic aperture radar (BiSAR) parametric imaging can reconstruct the structural
information of the target, which is one of the research hotspots for BiSAR imaging. However …
information of the target, which is one of the research hotspots for BiSAR imaging. However …
Data-driven compressed sensing for massive wireless access
The central challenge in massive machine-type communications (mMTC) is to connect a
large number of uncoordinated devices through a limited spectrum. The typical mMTC …
large number of uncoordinated devices through a limited spectrum. The typical mMTC …
MAda-Net: Model-adaptive deep learning imaging for SAR tomography
Y Wang, C Liu, R Zhu, M Liu, Z Ding… - IEEE Transactions on …, 2023 - ieeexplore.ieee.org
The compressive sensing (CS)-based tomographic SAR (TomoSAR) 3-D imaging method
has the shortcoming of low efficiency, mainly represented in two aspects: first, the CS solver …
has the shortcoming of low efficiency, mainly represented in two aspects: first, the CS solver …
Generalization error bounds for iterative recovery algorithms unfolded as neural networks
Motivated by the learned iterative soft thresholding algorithm (LISTA), we introduce a
general class of neural networks suitable for sparse reconstruction from few linear …
general class of neural networks suitable for sparse reconstruction from few linear …