arXiv subjects

Maryam Cheraghy

Publications and source records attributed to Maryam Cheraghy.

4 recordsLinked to original sources

arXiv2026

Resource Allocation and Beamforming in FIM-Assisted BS and STAR-BD-RIS-Aided NOMA: An AIW-Meta-Learning Approach

This paper investigates a flexible intelligent metasurface (FIM)-enabled wireless communication system that integrates simultaneously transmitting and reflecting beyond diagonal reconfigurable intelligent surfaces (STAR-BD-RIS) with non-orthogonal multiple access (NOMA). The considered system consists of a multi-antenna FIM-assisted base station (BS) supported by dual-sector BD-RIS. The FIM is composed of low-cost radiating elements capable of independent signal transmission and dynamic vertical reconfiguration (morphing). The objective is to maximize energy efficiency (EE) by jointly optimizing the BS beamforming, STAR-BD-RIS configuration, NOMA-related variables, and the FIM surface shape under practical power constraints. Due to the highly non-convex nature of the problem, an adaptive inverse-weighted Meta-Soft Actor-Critic (AIW-Meta-SAC) algorithm is proposed. Unlike conventional Meta-SAC approaches, the proposed method employs an adaptive weighting mechanism to effectively incorporate system constraints into the reward function, thereby improving learning efficiency and convergence behavior. Simulation results demonstrate that the proposed AIW-Meta-SAC significantly outperforms the Meta-DDPG baseline. Furthermore, the FIM-assisted STAR-BD-RIS architecture achieves notable energy efficiency gains compared to conventional benchmark schemes.

eess.SP↗

arXiv2026

UAV Trajectory Optimization via Improved Noisy Deep Q-Network

This paper proposes an Improved Noisy Deep Q-Network (Noisy DQN) to enhance the exploration and stability of Unmanned Aerial Vehicle (UAV) when applying deep reinforcement learning in simulated environments. This method enhances the exploration ability by combining the residual NoisyLinear layer with an adaptive noise scheduling mechanism, while improving training stability through smooth loss and soft target network updates. Experiments show that the proposed model achieves faster convergence and up to $+40$ higher rewards compared to standard DQN and quickly reach to the minimum number of steps required for the task 28 in the 15 * 15 grid navigation environment set up. The results show that our comprehensive improvements to the network structure of NoisyNet, exploration control, and training stability contribute to enhancing the efficiency and reliability of deep Q-learning.

eess.SY↗

arXiv2025

An Efficient Recommendation System in E-commerce using Passer learning optimization based on Bi-LSTM

Online reviews play a crucial role in shaping consumer decisions, especially in the context of e-commerce. However, the quality and reliability of these reviews can vary significantly. Some reviews contain misleading or unhelpful information, such as advertisements, fake content, or irrelevant details. These issues pose significant challenges for recommendation systems, which rely on user-generated reviews to provide personalized suggestions. This article introduces a recommendation system based on Passer Learning Optimization-enhanced Bi-LSTM classifier applicable to e-commerce recommendation systems with improved accuracy and efficiency compared to state-of-the-art models. It achieves as low as 1.24% MSE on the baby dataset. This lifts it as high as 88.58%. Besides, there is also robust performance of the system on digital music and patio lawn garden datasets at F1 of 88.46% and 92.51%, correspondingly. These results, made possible by advanced graph embedding for effective knowledge extraction and fine-tuning of classifier parameters, establish the suitability of the proposed model in various e-commerce environments.

cs.MM↗

arXiv2025

Hybrid FIM and STAR-BD-RIS-Aided Wireless Communications with Short Packet Length: A Meta-TD3 Approach

Reconfigurable intelligent surfaces (RIS) and flexible intelligent metasurfaces (FIM) have been widely adopted in multi-user wireless communication systems to enhance channel quality through simultaneous transmission and reflection of signals and three-dimensional reconfiguration of antennas. In this paper, we propose a novel system architecture that integrates the benefits of both technologies by deploying an FIM antenna at the base station (BS) and a simultaneously transmitting and reflecting beyond diagonal RIS (STAR-BD-RIS) along the transmission path to ensure sufficient received power for single-antenna users. The objective is to maximize the achievable sum rate considering the short block length by jointly optimizing the FIM surface configuration, the transmit beamforming vector, and STAR-BD-RIS phase shift matrix subject to practical constraints including minimum signal-to-interference-plus-noise ratio (SINR), power limitations, FIM constraint, and the STAR-BD-RIS phase-shift matrix. To solve the resulting non-convex optimization problem, we develop a learning-based approach that incorporates meta-learning into the twin delayed deep deterministic policy gradient (TD3) algorithm, referred to as Meta-TD3. The simulation results demonstrate that the proposed hybrid system outperforms conventional configurations employing either FIM or RIS alone, while the Meta-TD3 algorithm achieves superior performance compared to classic learning techniques.

eess.SP↗