Search

Your search keyword '"Ghalkha A"' showing total 6 results
Start Over Author "Ghalkha A"
6 results on '"Ghalkha A"'

1. Scalable and Resource-Efficient Second-Order Federated Learning via Over-the-Air Aggregation

Author

Ghalkha, Abdulmomen, Issaid, Chaouki Ben, and Bennis, Mehdi

Subjects
Computer Science - Machine Learning
Abstract

Second-order federated learning (FL) algorithms offer faster convergence than their first-order counterparts by leveraging curvature information. However, they are hindered by high computational and storage costs, particularly for large-scale models. Furthermore, the communication overhead associated with large models and digital transmission exacerbates these challenges, causing communication bottlenecks. In this work, we propose a scalable second-order FL algorithm using a sparse Hessian estimate and leveraging over-the-air aggregation, making it feasible for larger models. Our simulation results demonstrate more than $67\%$ of communication resources and energy savings compared to other first and second-order baselines., Comment: 5 pages, 1 figure, 4 subfigures, letter

Published
2024

6. Analysing random spatio-temporal variability of storm events for hydrological modelling

Author

Leonor Rodriguez-Sinobas, Sergio Zubelzu, Carlota Bernal, María Teresa Gómez, Jesús López Santiago, Andrea Zanella, Mehdi Bennis, Martina Capuzzo, Sara E. Matendo, Abdulmomen Ghalkha, and Chaouki Ben Issaid

Abstract

Events-based hydrology phenomena are affected by extreme spatio-temporal variability. Precipitation is the first source of variability. Storms can start at different times across a catchment and can evolve differently over time thus creating a complex frame for events-based hydrological modelling. Both are affected by a random character. On many occasions a scarce number of weather stations are available within the catchments so researchers and hydrologists are forced to use interpolation methods for estimating precipitation. Classical interpolation methods base on deterministic algorithms not properly accounting for the random character of storm. In this work we analyse the spatial variability of the recorded storms in a set of urban weather stations in Madrid delivering some criteria for dealing with spatio-temporal variability of storms for modelling events-based hydrological processes.

Published
2023
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results