Abstract
Identifying influential spreaders in temporal networks is crucial for understanding and controlling the dynamics of spreading. However, existing methods, such as temporal betweenness, closeness, pagerank, degree, and local path-based centrality, face several limitations, including high computational complexity, reliance on shortest paths, convergence issues, inability to capture influence dynamics with insufficient neighboring nodes, and a primary focus on local structural information. This paper presents PathSAGE, a novel method that addresses these problems. It integrates GraphSAGE, a deep learning model, to capture global node information while incorporating temporal local path counts as a key feature. Unlike other global feature-capturing methods, PathSAGE optimises computational complexity. Experimental results on thirteen real-world temporal networks demonstrate that PathSAGE outperforms the state-of-the-art methods in accurately identifying influential spreaders. PathSAGE exhibits a strong correlation with the Temporal Susceptible-Infected-Recovered (TSIR) model and achieves a relative improvement percentage (eta%) ranging from 0.12% to 70.70%. Additionally, PathSAGE attains the lowest average robustness value of 0.17, highlighting its effectiveness in identifying influential spreaders within temporal networks.

Abstract
The growing integration of renewable energy sources and the widespread electrification of the energy demand have significantly reduced the capacity margin of the electrical grid. This demands a more flexible approach to grid operation, for instance, combining real-time topology optimization and redispatching. Traditional expert-driven decision-making rules may become insufficient to manage the increasing complexity of real-time grid operations and derive remedial actions under the N-1 contingency. This work proposes a novel hybrid AI framework for power grid topology control that integrates genetic network programming (GNP), reinforcement learning, and decision trees. A new variant of GNP is introduced that is capable of evolving the decision-making rules by learning from data in a reinforcement learning framework. The graph-based evolutionary structure of GNP and decision trees enables transparent, traceable reasoning. The proposed method outperforms both a baseline expert system and a state-of-the-art deep reinforcement learning agent on the IEEE 118-bus system, achieving up to an 28% improvement in a key performance metric used in the Learning to Run a Power Network (L2RPN) competition.

Abstract
Textile-to-textile recycling is strongly constrained by upstream pre-processing, where post-consumer clothing must be identified, separated, and prepared under high variability in materials, appearance, and contamination. This paper presents a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)-guided systematic literature review of intelligent and automated technologies for textile recycling pre-processing covering the interval between 2015 to 2025. After screening and quality assessment, 21 primary studies published between 2020 and 2025 were included. The literature is synthesized across three task families: (i) identificationof fiber/material, composition, or color; (ii) sorting, considered only when explicit separation strategies are defined to operationalize identification outcomes into routing actions or output streams; and (iii) contaminant detection and/or removal, targeting non-recyclable items. Results show that identification dominates the field (19/21 studies), supported by Red-Green-Blue (RGB) and red-green-blue plus depth (RGB-D) imaging and material-signature sensing, including near-infrared (NIR) spectroscopy, hyperspectral imaging (HSI), and Raman spectroscopy. In contrast, sorting as a defined separation stage is less frequent (4/21), and contaminant-related automation remains sparse (3/21). Most studies are validated in laboratory conditions, with limited semi-industrial evidence, highlighting a persistent perception-to-action gap. Overall, the review indicates that robust separation strategies, representative datasets, and end-to-end system integration remain key bottlenecks for scalable automated textile recycling pre-processing.

Abstract
The Industry 5.0 concept guides the industry to the premise of sustainability, resilience and human-centric solutions. The last related pillar tries to create solutions to empower the people in production line processes since solutions should be designed to be easy to use and easy to learn without discarding the working people. In this regard, it's natural that robots become closer to humans in industrial applications where it is possible to absorb human-machine qualities. Robotic grasping has widespread application with a wide range of applicability. However, engineers and shop-floor operators spend time finding a fast response solution when the production demand changes. Aiming to create a tool to help this procedure in a human-centred fashion, the current paper proposes a programming-by-demonstration solution that is easy to use, reuse, adapt, and increment with its modular design.

Abstract
Hydrogen (H2) detection has become extremely important in recent years due to the increasing need for sustainable alternative energy sources. In this field, optical sensors can contribute significantly due to remote interrogation capabilities and the absence of ignition sources. Among the different H2 optical sensors, plasmonic sensors appear to be a very sensitive technology; however, they require expensive plasmonic materials like gold or silver, which, together with a palladium-sensitive layer, can increase the sensor cost. In addition, plasmonic bands are usually outside the ideal infrared range for remote interrogation, between 1500 and 1600 nm. This work presents a polymer-protected Tamm Plasmon Resonance (TPR) sensor with a well-defined resonance band at 1572 nm composed of SiO2, TiO2 layers, and palladium as a sensitive layer. This architecture can reduce the production cost of sensing structures, replacing plasmonic films with dielectric materials, while offering improved resonance definition at longer wavelengths. First, numerical calculations were carried out using the Transfer-Matrix Method to study the impact of the thickness of each layer, incidence angle, and light polarization on the resonance band wavelength and H2 sensitivity. The optimized structure was then fabricated, exhibiting a wavelength shift of 9.5 nm to 4 vol % of H2, a response time of 30 s, and no cross-sensitivity to methane or ammonia. The sensor also demonstrated high stability and resistance to environmental degradation up to eight days. These results emphasize the advantages of TPR structures for gas sensing in the infrared spectral range, opening new avenues for remote plasmonic sensing. © 2026 The Authors. Published by American Chemical Society

Abstract
Eye diseases can have highly adverse outcomes without an early diagnosis and correct monitoring. Retinopathy of Prematurity (ROP) Plus Form, in particular, is a disease that can lead to childhood blindness, and its diagnosis requires medical experts to examine the retinal condition manually. Although developments in screening equipment have helped, this is still a time-consuming and subjective task. The development of automatic tools for Retinal Blood Vessel Segmentation allows the extraction of blood vessels from fundus images, which healthcare experts can then use to perform the diagnosis, monitoring, and prognosis of eye diseases. Thus, developing such a segmentation tool is a widely explored task with different methodologies that can be followed. However, many studies try to segment all the blood vessels rather than only the most important ones. In this work, we present a segmentation pipeline to segment only the main vessels whose characteristics can be used to assess ROP Plus Form disease. This pipeline uses different operations and filters, including CIELAB Enhancement, Background Normalization, Bell-Shaped Gaussian Matched Filtering, Modified Top-Hat operation, and Frangi Filtering. The final segmentation is done by determining a threshold value using the Triangle Threshold algorithm. The pipeline was tested in the well-known DRIVE Database, achieving an Accuracy of 0.949, Specificity of 0.963, and Sensitivity of 0.756.