Thanh V. Pham

Trinh K. Nguyen, Hung K. Hoang, Thanh V. Pham et al.

This paper investigates the secrecy sum-rate (SSR) performance of optical intelligent reflecting surface (OIRS)-assisted multi-user visible light communication (VLC) systems under line-of-sight (LoS) blockages. To mitigate physical obstructions and internal eavesdropping, a joint optimization problem is formulated to maximize the SSR through the co-design of the transmission precoder and OIRS units assignment. Due to the binary constraints and coupled variables, the problem is highly non-convex. To solve it efficiently, an alternating optimization (AO) framework integrating the concave-convex procedure (CCCP) and first-order Taylor approximations is developed. Simulation results demonstrate the convergence of the proposed algorithm and show that increasing the number of OIRS reflecting units yields significant SSR gains.

Hung K. Hoang, Chuyen T. Nguyen, Thang K. Nguyen et al.

This paper investigates the secrecy sum rate (SSR) of rate-splitting multiple access (RSMA)-based visible light communication (VLC) systems considering internal eavesdropping, where legitimate users may intercept private data intended for others. We formulate an optimization problem to maximize the SSR of the system, which is inherently non-convex due to the complex coupling of the objective function and constraints. To this end, two different approaches based on the convex-concave procedure (CCCP) and semidefinite relaxation (SDR) are leveraged to solve the non-convex parameterized problem. A central focus of this work is the investigation of channel similarity (CS), which serves as a metric for quantifying spatial correlation, and its impact on SSR performance. To mitigate the performance degradation caused by high spatial correlation, we propose a channel similarity reduction (CSR) clustering strategy that proactively minimizes CS to restore the system's degrees of freedom (DoF). Numerical results are provided to demonstrate the performance of the two proposed algorithms under various levels of CS. More importantly, the findings reveal that our proposed CSR-clustering strategy significantly outperforms existing baselines, effectively overcoming the secrecy performance ceiling caused by high spatial correlation.

Tuan A. Hoang, Phuc V. Trinh, Shinya Sugiura et al.

This paper investigates near-field propagation in optical reconfigurable intelligent surface (ORIS)-assisted free-space optical (FSO) communication systems. Unlike conventional far-field scenarios, near-field propagation involves complex diffraction effects that hinder tractable closed-form analysis. To address this issue, a numerical framework for evaluating the optical field distribution of ORIS-assisted FSO links is proposed. Specifically, two numerical approaches are considered: direct Riemann-sum evaluation and a fast Fourier transform (FFT)-based method. Although the Riemann sum approach provides accurate field estimation, it incurs extremely high computational complexity due to the fine spatial discretization of the ORIS surface required at optical wavelengths. To improve computational efficiency, the optical-field calculation is reformulated as a convolution in the spatial-frequency domain, enabling efficient FFT-based propagation analysis. Simulation results demonstrate that the proposed FFT-based method achieves accuracy comparable to that of the Riemann-sum approach while significantly reducing computational complexity.

James Jin Kang, Dang Bui, Thanh Pham et al.

The growing use of large language models in sensitive domains has exposed a critical weakness: the inability to ensure that private information can be permanently forgotten. Yet these systems still lack reliable mechanisms to guarantee that sensitive information can be permanently removed once it has been used. Retraining from the beginning is prohibitively costly, and existing unlearning methods remain fragmented, difficult to verify, and often vulnerable to recovery. This paper surveys recent research on machine unlearning for LLMs and considers how far current approaches can address these challenges. We review methods for evaluating whether forgetting has occurred, the resilience of unlearned models against adversarial attacks, and mechanisms that can support user trust when model complexity or proprietary limits restrict transparency. Technical solutions such as differential privacy, homomorphic encryption, federated learning, and ephemeral memory are examined alongside institutional safeguards including auditing practices and regulatory frameworks. The review finds steady progress, but robust and verifiable unlearning is still unresolved. Efficient techniques that avoid costly retraining, stronger defenses against adversarial recovery, and governance structures that reinforce accountability are needed if LLMs are to be deployed safely in sensitive applications. By integrating technical and organizational perspectives, this study outlines a pathway toward AI systems that can be required to forget, while maintaining both privacy and public trust.