Deep Proximal Unrolling: Algorithmic Framework, Convergence Analysis and Applications

Liu, Risheng; Cheng, Shuying; Ma, Long; Fan, Xin; Luo, Zhongxuan

doi:10.1109/tip.2019.2913536

Cited by 56 publications

(21 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Algorithm unfolding [11,[24][25][26] refers to the general notion of building a problem-specific neural architecture with layers inspired by an iterative solution to the same problem. In this paper, we develop and evaluate a way to unfold the classical WMMSE [5] method using graph neural networks for power allocation in wireless networks.…”

Section: Discussionmentioning

confidence: 99%

Efficient Power Allocation Using Graph Neural Networks and Deep Algorithm Unfolding

Chowdhury

Verma²,

Rao³

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

We study the problem of optimal power allocation in a single-hop ad hoc wireless network. In solving this problem, we propose a hybrid neural architecture inspired by the algorithmic unfolding of the iterative weighted minimum mean squared error (WMMSE) method, that we denote as unfolded WMMSE (UWMMSE). The learnable weights within UWMMSE are parameterized using graph neural networks (GNNs), where the time-varying underlying graphs are given by the fading interference coefficients in the wireless network. These GNNs are trained through a gradient descent approach based on multiple instances of the power allocation problem. Once trained, UWMMSE achieves performance comparable to that of WMMSE while significantly reducing the computational complexity. This phenomenon is illustrated through numerical experiments along with the robustness and generalization to wireless networks of different densities and sizes.

show abstract

Section: Discussionmentioning

confidence: 99%

Efficient Power Allocation Using Graph Neural Networks and Deep Algorithm Unfolding

Chowdhury

Verma²,

Rao³

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…This is a key advantage to unrolled networks, as their origins have well‐accepted and well‐defined optimisation models with guaranteed convergence. However, theoretical convergence may no longer be valid due to the dynamic nature of network parameters 53 . Implementation of unrolled networks also typically necessitates the evaluation of signal operators such as the Fourier transform, both in the training and execution phases.…”

Section: Compressed Sensingmentioning

confidence: 99%

“…However, theoretical convergence may no longer be valid due to the dynamic nature of network parameters. 53 Implementation of unrolled networks also typically necessitates the evaluation of signal operators such as the Fourier transform, both in the training and execution phases. This can significantly increase processing time and memory requirements, particularly when fast GPU implementations are not available.…”

Section: Compressed Sensingmentioning

confidence: 99%

Deep learning in magnetic resonance image reconstruction

et al. 2021

View full text Add to dashboard Cite

Summary Magnetic resonance (MR) imaging visualises soft tissue contrast in exquisite detail without harmful ionising radiation. In this work, we provide a state‐of‐the‐art review on the use of deep learning in MR image reconstruction from different image acquisition types involving compressed sensing techniques, parallel image acquisition and multi‐contrast imaging. Publications with deep learning‐based image reconstruction for MR imaging were identified from the literature (PubMed and Google Scholar), and a comprehensive description of each of the works was provided. A detailed comparison that highlights the differences, the data used and the performance of each of these works were also made. A discussion of the potential use cases for each of these methods is provided. The sparse image reconstruction methods were found to be most popular in using deep learning for improved performance, accelerating acquisitions by around 4–8 times. Multi‐contrast image reconstruction methods rely on at least one pre‐acquired image, but can achieve 16‐fold, and even up to 32‐ to 50‐fold acceleration depending on the set‐up. Parallel imaging provides frameworks to be integrated in many of these methods for additional speed‐up potential. The successful use of compressed sensing techniques and multi‐contrast imaging with deep learning and parallel acquisition methods could yield significant MR acquisition speed‐ups within clinical routines in the near future.

show abstract

“…They proposed learned ISTA (LISTA), based on unrolling the update steps of ISTA [14] into layers of a feedforward neural network. Subsequent studies have demonstrated the efficacy of this class of modelbased deep learning architectures [33] in compressed sensing and sparse-recovery applications [32], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43]. Deep-unfolding combines the advantages of both data-driven and iterative techniques.…”

Section: A Prior Artmentioning

confidence: 99%

DuRIN: A Deep-unfolded Sparse Seismic Reflectivity Inversion Network

Mache,

Pokala,

Rajendran

et al. 2021

Preprint

View full text Add to dashboard Cite

We consider the reflection seismology problem of recovering the locations of interfaces and the amplitudes of reflection coefficients from seismic data, which are vital for estimating the subsurface structure. The reflectivity inversion problem is typically solved using greedy algorithms and iterative techniques. Sparse Bayesian learning framework, and more recently, deep learning techniques have shown the potential of data-driven approaches to solve the problem. In this paper, we propose a weighted minimax-concave penalty-regularized reflectivity inversion formulation and solve it through a modelbased neural network. The network is referred to as deepunfolded reflectivity inversion network (DuRIN). We demonstrate the efficacy of the proposed approach over the benchmark techniques by testing on synthetic 1-D seismic traces and 2-D wedge models and validation with the simulated 2-D Marmousi2 model and real data from the Penobscot 3D survey off the coast of

show abstract

Deep Proximal Unrolling: Algorithmic Framework, Convergence Analysis and Applications

Cited by 56 publications

References 40 publications

Efficient Power Allocation Using Graph Neural Networks and Deep Algorithm Unfolding

Efficient Power Allocation Using Graph Neural Networks and Deep Algorithm Unfolding

Deep learning in magnetic resonance image reconstruction

DuRIN: A Deep-unfolded Sparse Seismic Reflectivity Inversion Network

Contact Info

Product

Resources

About