2021 IEEE International Conference on Acoustics, Speech and Signal Processing
6-11 June 2021 • Toronto, Ontario, Canada
Extracting Knowledge from Information
| Paper ID | SPCOM-1.3 | ||
| Paper Title | EFFICIENT POWER ALLOCATION USING GRAPH NEURAL NETWORKS AND DEEP ALGORITHM UNFOLDING | ||
| Authors | Arindam Chowdhury, Rice University, United States; Gunjan Verma, Chirag Rao, Ananthram Swami, US Army’s CCDC Army Research Laboratory, United States; Santiago Segarra, Rice University, United States | ||
| Session | SPCOM-1: Signal Processing for Networks | ||
| Location | Gather.Town | ||
| Session Time: | Tuesday, 08 June, 16:30 - 17:15 | ||
| Presentation Time: | Tuesday, 08 June, 16:30 - 17:15 | ||
| Presentation | Poster | ||
| Topic | Signal Processing for Communications and Networking: [SPCN-NETW] Networks and Network Resource allocation | ||
| IEEE Xplore Open Preview | Click here to view in IEEE Xplore | ||
| Abstract | We study the problem of optimal power allocation in a single-hop ad hoc wireless network. In solving this problem, we propose a hybrid neural architecture inspired by the algorithmic unfolding of the iterative weighted minimum mean squared error (WMMSE) method, that we denote as unfolded WMMSE (UWMMSE). The learnable weights within UWMMSE are parameterized using graph neural networks (GNNs), where the time-varying underlying graphs are given by the fading interference coefficients in the wireless net-work. These GNNs are trained through a gradient descent approach based on multiple instances of the power allocation problem. Once trained, UWMMSE achieves performance comparable to that of WMMSE while significantly reducing the computational complexity. This phenomenon is illustrated through numerical experiments along with the robustness and generalization to wireless networks of different densities and sizes | ||