http://hdl.handle.net/10071/5659 2026-04-04T00:40:19Z http://hdl.handle.net/10071/36504 Título próprio: Evaluating SDG network models: A network science ontology-based framework Autoria: Pasishnyk, N.; Lopes, R. J. Resumo: With only 18% of Sustainable Development Goals (SDGs) on track for 2030, systems-based approaches to understanding their interdependencies are essential. Network science can reveal leverage points and guide prioritisation, yet it is often applied without sufficient domain integration, obscuring rather than clarifying sustainability dynamics. We present an eight-step framework for evaluating network science applications in SDG research. This framework was applied to 25 studies selected via a scoping review process focused on SDG interactions. Using the proposed framework each paper was coded and classified into A/B/C methodological tiers. The analysis reveals two dominant patterns: semantic/expert-based approaches (11 studies) and indicator/statistical approaches (12 studies). Beyond these, one study implements a multiplex design and another a heterogeneous multilayer architecture. Critically, 96% of these papers focus on formal SDG structures rather than the actors, processes, and mechanisms through which targets are achieved, limiting practical utility. The framework makes explicit how modelling choices encode theoretical assumptions and supports like-with-like comparison, meta-analysis and evidence synthesis. As AI-enabled knowledge synthesis proliferates, such transparency steers SDG modelling toward implementation-relevant representations that preserve contextual factors shaping real-world transformations. 2026-01-01T00:00:00Z http://hdl.handle.net/10071/36408 Título próprio: Optimizing the achievable sum-rate in OFDM-based Multi-User MIMO systems assisted by multiple Beyond-Diagonal RISs Autoria: Mendes, D.; Souto, N.; Pavia, J. P.; Silva, J. Resumo: Massive multiple-input, multiple-output (MIMO) systems operating in the millimeter wave (mmWave) and terahertz (THz) frequency bands offer high data rates and spatial multiplexing, yet they face significant propagation challenges. Reconfigurable intelligent surfaces (RISs) and their different architectures have emerged as a promising solution to these challenges, having the potential to enhance system performance. This paper addresses the joint sum-rate maximization problem in multi-user, multi-stream, multi-carrier MIMO systems aided by multiple parallel RIS panels. To minimize the inter-user interference, we adopt a problem formulation that adds a regularization term. To solve the resulting problem, we then propose a regularized cyclic block proximal gradient (MU-RCBPG) algorithm, which can jointly optimize precoders and RIS phase shifts without increasing the complexity compared to traditional singl-evalued decomposition (SVD)-based methods. The resulting algorithm has a flexible design that allows it to support configurations with beyond-diagonal RIS (BD-RIS), conventional diagonal RIS (D-RIS), and active D-RIS. Numerical results demonstrate that the MU-RCBPG algorithm outperforms existing RIS-aided schemes in various scenarios. 2026-01-01T00:00:00Z http://hdl.handle.net/10071/36397 Título próprio: EcDiff-LLIE: Event-conditional diffusion model for structure-preserving low-light image enhancement Autoria: Maqsood, R.; Nunes, P.; Soares, L. D.; Conti, C. Resumo: Low-light image enhancement (LLIE) aims to restore the visual quality of poorly illuminated images by recovering fine details and textures while suppressing noise and artifacts. Recently, diffusion models have shown superior generative capabilities for LLIE. However, existing diffusion-based methods condition the denoising process only on low-light images or features derived from them (e.g., structural or illumination maps). Since the low-light images are severely degraded, this limits the denoising model’s ability to restore fine structure and reduce artifacts. In this work, we show that the event data captured simultaneously with the low-light images provides complementary high-dynamic-range and high-temporal-resolution structural information that can overcome this limitation. Therefore, we propose EcDiff-LLIE, a novel event-conditional diffusion framework for LLIE. At its core, we introduce a multimodality denoising network that conditions on both low-light images and concurrent event streams. To effectively fuse the two modalities, we design a cross-modality attention block that bridge their domain differences, while also enabling long-range dependency modeling for improved structural preservation. Experiments on the synthetic SDSD and real-world SDE datasets show significant improvements in quantitative evaluation metrics. Furthermore, evaluation on the high-resolution real-world HUE dataset further shows the generalization ability of the proposed framework. 2026-01-01T00:00:00Z http://hdl.handle.net/10071/36384 Título próprio: Beamforming optimization and system level assessment in RIS-aided MIMO systems comprising hybrid precoding architectures Autoria: Mendes, D.; Pavia, J. P.; Souto, N.; Silva, J.; Correia, A. Resumo: The terahertz (THz) band is a candidate technology for future sixth-generation (6G) wireless networks that could support increasingly demanding requirements, such as high wireless traffic volumes and transmission rates. However, the limited range and high propagation losses at these frequencies present several challenges that must be overcome. One emerging solution is the use of reconfigurable intelligent surfaces (RIS), which optimize communication network performance in combination with ultra-massive multiple-input multiple-output (UM-MIMO) antennas. UM-MIMO’s large number of antennas provides highly directional beams, enabling reliable data propagation from the transmitter to the receiver at THz frequencies. However, it can substantially increase implementation complexity. This paper proposes a joint hybrid precoder and RIS optimization algorithm to overcome these challenges. The algorithm is designed to maximize the achievable rate of THz UM-MIMO communications, by segregating digital and analog precoder computations and adopting hybrid architectures: fully connected (FC), array-of-subarrays (AoSA), and dynamic array-of-subarrays (DAoSA). The proposed algorithm supports multicarrier transmission and the use of multiple, parallel RIS panels deployed throughout the propagation path. Numerical simulations demonstrate the efficiency and versatility of the algorithm, particularly in contexts where THz systems operate under severe constraints. System-level simulations in a 300 GHz office environment reveal that distributing multiple parallel RIS panels throughout the environment yields the maximum achievable throughput. RIS deployment offers the greatest coverage and throughput gains in low-density scenarios but provides diminishing returns as density increases. 2026-01-01T00:00:00Z