Embedded Files
The problem of interference management or resource allocation (i.e., wireless link scheduling, power allocation) can be learned through deep learning (DL) models. However, they require large training samples. In this series of works, we aim to reduce the number of training samples, while approaching the sum-rate as that of the state-of-the-art DL models. The following graph outlines the consecutive stages of the proposed work.
Wireless Link Scheduling over Riemannian manifolds:
This work focuses on using machine learning models over Riemannian manifolds to reduce the learning samples. Supervised and unsupervised ML models have been proposed. For more information, please refer to the journal publication below.
R. Shelim and A. S. Ibrahim, "Geometric Machine Learning Over Riemannian Manifolds for Wireless Link Scheduling," in IEEE Access, vol. 10, pp. 22854-22864, Feb. 2022, doi: 10.1109/ACCESS.2022.3153324.
2. Wireless Link Scheduling over Recurrent Riemannian manifolds
This work employs spatio-temporal correlation of dynamic networks to reduce the number of training samples. For more information, please refer to the journal publication below.
R. Shelim and A. S. Ibrahim, "Wireless Link Scheduling Over Recurrent Riemannian Manifolds," in IEEE Transactions on Vehicular Technology, doi: 10.1109/TVT.2022.3228212.
3. Learning Wireless Power Allocation through Graph Convolutional Regression Networks over Riemannian Manifolds
This work develops a learning model for power allocation with few training samples, which is vital in dynamic networks (e.g., vehicular networks). The proposed model transforms Euclidean-based network layouts and power allocation problems into Riemannian (i.e., non-Euclidean) manifolds, which is shown to require fewer learning parameters and hence shorter learning time. Such transformation is possible thanks to the symmetric positive definite (SPD) property of spectral representation (i.e., Laplacian matrix) of network layouts. In particular, we propose a graph convolutional regression network (GCRN) for predicting power allocation over Riemannian manifolds in an unsupervised manner.
For more information, please refer to the journal publication below.
R. Shelim and A. S. Ibrahim, "Learning Wireless Power Allocation Through Graph Convolutional Regression Networks Over Riemannian Manifolds," in IEEE Transactions on Vehicular Technology, doi: 10.1109/TVT.2023.3325200.
Page updated
Google Sites
Report abuse