Hard Thresholding Pursuit: An Algorithm for Compressive Sensing

By Simon Foucart. SIAM Journal on Numerical Analysis, 2011

549 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Computational Mechanics, Computer Networks and Communications

View Paper: https://www.semanticscholar.org/paper/e10a650f8c557c64b0a4554c529eead94ad431ed

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

1. Oluwasanmi Koyejo
2. Jacky Y. Zhang
3. Raajay Viswanathan
4. Ashish Kapoor
5. Anastasios Kyrillidis
6. Purushottam Kar
7. Rekha Khanna
8. Prateek Jain

In Computational Mechanics

1. Purushottam Kar
2. Anastasios Kyrillidis
3. Oluwasanmi Koyejo
4. Jacky Y. Zhang
5. Raajay Viswanathan
6. Ashish Kapoor
7. Rekha Khanna
8. Prateek Jain

In Computer Networks and Communications

1. Oluwasanmi Koyejo
2. Raajay Viswanathan
3. Ashish Kapoor
4. Anastasios Kyrillidis
5. Purushottam Kar
6. Prateek Jain

Citations in the corpus

• Multilabel Classification using Bayesian Compressed Sensing (2012, methodology): “For test points, the learnt regression function is applied in the reduced space and then standard recovery algorithms from CS literature are used to obtain sparse predicted labels [8, 9].”
• Non-convex Optimization for Machine Learning (2017, methodology): “The IHT method is a part of a larger class of hard thresholding techniques, which include algorithms such as Iterative Hard Thresholding (IHT) [Blumensath, 2011], Gradient Descent with Sparsification (GraDeS) [Garg and Khandekar, 2009], and Hard Thresholding Pursuit (HTP) [Foucart, 2011].”
• Bayesian Coresets: Revisiting the Nonconvex Optimization Perspective (2021, background): “For example, in (Blumensath & Davies, 2009; Foucart, 2011) strong isometry constant bounds are assumed to allow step size µ = 1 for all the iterations, and thus remove the requirement of hyper-parameter tuning.”