-
Efficient Federated Low Rank Matrix Completion
Authors:
Ahmed Ali Abbasi,
Namrata Vaswani
Abstract:
In this work, we develop and analyze a Gradient Descent (GD) based solution, called Alternating GD and Minimization (AltGDmin), for efficiently solving the low rank matrix completion (LRMC) in a federated setting. LRMC involves recovering an $n \times q$ rank-$r$ matrix $\Xstar$ from a subset of its entries when $r \ll \min(n,q)$. Our theoretical guarantees (iteration and sample complexity bounds)…
▽ More
In this work, we develop and analyze a Gradient Descent (GD) based solution, called Alternating GD and Minimization (AltGDmin), for efficiently solving the low rank matrix completion (LRMC) in a federated setting. LRMC involves recovering an $n \times q$ rank-$r$ matrix $\Xstar$ from a subset of its entries when $r \ll \min(n,q)$. Our theoretical guarantees (iteration and sample complexity bounds) imply that AltGDmin is the most communication-efficient solution in a federated setting, is one of the fastest, and has the second best sample complexity among all iterative solutions to LRMC. In addition, we also prove two important corollaries. (a) We provide a guarantee for AltGDmin for solving the noisy LRMC problem. (b) We show how our lemmas can be used to provide an improved sample complexity guarantee for AltMin, which is the fastest centralized solution.
△ Less
Submitted 10 May, 2024;
originally announced May 2024.
-
r-local sensing: Improved algorithm and applications
Authors:
Ahmed Ali Abbasi,
Abiy Tasissa,
Shuchin Aeron
Abstract:
The unlabeled sensing problem is to solve a noisy linear system of equations under unknown permutation of the measurements. We study a particular case of the problem where the permutations are restricted to be r-local, i.e. the permutation matrix is block diagonal with r x r blocks. Assuming a Gaussian measurement matrix, we argue that the r-local permutation model is more challenging compared to…
▽ More
The unlabeled sensing problem is to solve a noisy linear system of equations under unknown permutation of the measurements. We study a particular case of the problem where the permutations are restricted to be r-local, i.e. the permutation matrix is block diagonal with r x r blocks. Assuming a Gaussian measurement matrix, we argue that the r-local permutation model is more challenging compared to a recent sparse permutation model. We propose a proximal alternating minimization algorithm for the general unlabeled sensing problem that provably converges to a first order stationary point. Applied to the r-local model, we show that the resulting algorithm is efficient. We validate the algorithm on synthetic and real datasets. We also formulate the 1-d unassigned distance geometry problem as an unlabeled sensing problem with a structured measurement matrix.
△ Less
Submitted 14 February, 2022; v1 submitted 26 October, 2021;
originally announced October 2021.
-
A Mobile Cloud-Based eHealth Scheme
Authors:
Yihe Liu,
Aaqif Afzaal Abbasi,
Atefeh Aghaei,
Almas Abbasi,
Amir Mosavi,
Shahab Shamshirband,
Mohammed A. A. Al-qaness
Abstract:
Mobile cloud computing is an emerging field that is gaining popularity across borders at a rapid pace. Similarly, the field of health informatics is also considered as an extremely important field. This work observes the collaboration between these two fields to solve the traditional problem of extracting Electrocardiogram signals from trace reports and then performing analysis. The developed syst…
▽ More
Mobile cloud computing is an emerging field that is gaining popularity across borders at a rapid pace. Similarly, the field of health informatics is also considered as an extremely important field. This work observes the collaboration between these two fields to solve the traditional problem of extracting Electrocardiogram signals from trace reports and then performing analysis. The developed system has two front ends, the first dedicated for the user to perform the photographing of the trace report. Once the photographing is complete, mobile computing is used to extract the signal. Once the signal is extracted, it is uploaded into the server and further analysis is performed on the signal in the cloud. Once this is done, the second interface, intended for the use of the physician, can download and view the trace from the cloud. The data is securely held using a password-based authentication method. The system presented here is one of the first attempts at delivering the total solution, and after further upgrades, it will be possible to deploy the system in a commercial setting.
△ Less
Submitted 15 April, 2020;
originally announced April 2020.
