Learned ISTA with Error-based Thresholding for Adaptive Sparse Coding

12/21/2021
by   Ziang Li, et al.
0

The learned iterative shrinkage thresholding algorithm (LISTA) introduces deep unfolding models with learnable thresholds in some shrinkage functions for sparse coding. Drawing on some theoretical insights, we advocate an error-based thresholding (EBT) mechanism for LISTA, which leverages a function of the layer-wise reconstruction error to suggest an appropriate threshold value for each observation on each layer. We show that the EBT mechanism well disentangles the learnable parameters in the shrinkage functions from the reconstruction errors, making them more adaptive to the various observations. With rigorous theoretical analyses, we show that the proposed EBT can lead to a faster convergence on the basis of LISTA and its variants, in addition to its higher adaptivity. Extensive experimental results confirm our theoretical analyses and verify the effectiveness of our methods.

READ FULL TEXT
research
06/22/2021

Learned Interpretable Residual Extragradient ISTA for Sparse Coding

Recently, the study on learned iterative shrinkage thresholding algorith...
research
02/06/2023

RDFNet: Regional Dynamic FISTA-Net for Spectral Snapshot Compressive Imaging

Deep convolutional neural networks have recently shown promising results...
research
01/30/2021

Metalearning: Sparse Variable-Structure Automata

Dimension of the encoder output (i.e., the code layer) in an autoencoder...
research
12/15/2015

Learning optimal nonlinearities for iterative thresholding algorithms

Iterative shrinkage/thresholding algorithm (ISTA) is a well-studied meth...
research
10/05/2020

Neurally Augmented ALISTA

It is well-established that many iterative sparse reconstruction algorit...
research
06/20/2023

Representation Sparsification with Hybrid Thresholding for Fast SPLADE-based Document Retrieval

Learned sparse document representations using a transformer-based neural...

Please sign up or login with your details

Forgot password? Click here to reset