Showing 1–2 of 2 results for author: Bui, V P
-
Goal-Oriented Communications in Federated Learning via Feedback on Risk-Averse Participation
Authors:
Shashi Raj Pandey,
Van Phuc Bui,
Petar Popovski
Abstract:
We treat the problem of client selection in a Federated Learning (FL) setup, where the learning objective and the local incentives of the participants are used to formulate a goal-oriented communication problem. Specifically, we incorporate the risk-averse nature of participants and obtain a communication-efficient on-device performance, while relying on feedback from the Parameter Server (\texttt…
▽ More
We treat the problem of client selection in a Federated Learning (FL) setup, where the learning objective and the local incentives of the participants are used to formulate a goal-oriented communication problem. Specifically, we incorporate the risk-averse nature of participants and obtain a communication-efficient on-device performance, while relying on feedback from the Parameter Server (\texttt{PS}). A client has to decide its transmission plan on when not to participate in FL. This is based on its intrinsic incentive, which is the value of the trained global model upon participation by this client. Poor updates not only plunge the performance of the global model with added communication cost but also propagate the loss in performance on other participating devices. We cast the relevance of local updates as \emph{semantic information} for develo** local transmission strategies, i.e., making a decision on when to ``not transmit". The devices use feedback about the state of the PS and evaluate their contributions in training the learning model in each aggregation period, which eventually lowers the number of occupied connections. Simulation results validate the efficacy of our proposed approach, with up to $1.4\times$ gain in communication links utilization as compared with the baselines.
△ Less
Submitted 19 May, 2023;
originally announced May 2023.
-
Joint Beam Placement and Load Balancing Optimization for Non-Geostationary Satellite Systems
Authors:
Van Phuc Bui,
Trinh Van Chien,
Eva Lagunas,
Joël Grotz,
Symeon Chatzinotas,
Björn Ottersten
Abstract:
Non-geostationary (Non-GSO) satellite constellations have emerged as a promising solution to enable ubiquitous high-speed low-latency broadband services by generating multiple spot-beams placed on the ground according to the user locations. However, there is an inherent trade-off between the number of active beams and the complexity of generating a large number of beams. This paper formulates and…
▽ More
Non-geostationary (Non-GSO) satellite constellations have emerged as a promising solution to enable ubiquitous high-speed low-latency broadband services by generating multiple spot-beams placed on the ground according to the user locations. However, there is an inherent trade-off between the number of active beams and the complexity of generating a large number of beams. This paper formulates and solves a joint beam placement and load balancing problem to carefully optimize the satellite beam and enhance the link budgets with a minimal number of active beams. We propose a two-stage algorithm design to overcome the combinatorial structure of the considered optimization problem providing a solution in polynomial time. The first stage minimizes the number of active beams, while the second stage performs a load balancing to distribute users in the coverage area of the active beams. Numerical results confirm the benefits of the proposed methodology both in carrier-to-noise ratio and multiplexed users per beam over other benchmarks.
△ Less
Submitted 29 July, 2022;
originally announced July 2022.
