Abstract
This paper introduces PRISMA, a hybrid multi-agent Deep Reinforcement Learning (DRL) framework for solving the Flexible Job-shop Scheduling Problem (FJSP). It uses Constraint Programming (CP) solutions to pretrain decentralized policies and to guide exploration during training. Although DRL can generate fast solutions for large combinatorial problems, it often fails to match the quality of optimization methods, motivating the integration with hybrid frameworks. The growing interest in embedding domain knowledge into learning algorithms has produced several hybrid formulations, yet their potential remains underexplored, particularly in multi-agent settings. PRISMA combines supervised and reinforcement learning within a multi-agent framework, where CP solutions are used to (i) learn expert decisions through imitation learning, and (ii) train an auxiliary network that guides DRL training via reward shaping. A shared graph network is adopted for transferring system-level knowledge into machine-level observations, enabling fast and consistent inference from enriched local embed-dings. To the best of our knowledge, PRISMA introduces the first expert-derived guidance mechanism for the FJSP and is among the earliest to apply imitation learning within a multi-agent formulation. By combining both modules, it strengthens the bridge between optimization and learning-based methods, where such dual integrations remain scarce. Experimental results show faster convergence and higher solution quality than state-ofthe-art DRL models. PRISMA achieves an average optimality gap of 6.74%, corresponding to a 50% relative improvement over the single-agent baseline, while reducing inference time. These findings reinforce the value of merging optimization accuracy with the flexibility of multi-agent DRL for efficient scheduling.

Abstract
This study applies phase plane analysis of medio-lateral, anteroposterior, and vertical directions for the coordination assessment of whole-body (WB) center of mass (COM) movement during the impulse phase of a standard maximum vertical jump (MVJ) with long, short, and no countermovement (CM). A video system and force platform were used, with the amplitudes of WB COM excursion obtained from image-based motion capture at each anatomical direction, and the 2D and 3D mean radial distance were compared under long, short, and no CM conditions. The estimate of the population mean length was used as a measure of distribution concentration, and the Rayleigh statistical test for circular data was applied with the sample distribution critical value. Watson's U2 goodness-of-fit test for the von Mises distribution was used with the mean direction and concentration factor. The applied metrics led to the detection of shared and specific features in the global and phase plane analysis of WB COM movement coordination in the medio-lateral, anteroposterior, and vertical directions during long, short, and no CM conditions in relation to MVJ performance assessed from ground reaction force (GRF) through the force platform. Thus, long, short, and no CM impulses share lower amplitudes of WB COM excursion in the medio-lateral direction and mean radial distance to its mean, whereas the anteroposterior and vertical excursion of WB COM, along with the 2D transversal and 3D spatial length of the WB COM path, present as potential predictors of MVJ performance, with distinct behavior in long CM compared to short and no CM. Additionally, the applied workflow on generalized phase plane analysis led to the detection, through complementary metrics, of the anatomical WB COM movement directions with higher coordination based on phase concentration tests at 5% significance, in line with MVJ performance under different CM conditions.

Abstract
This paper proposes an optimized Reconfigurable Internet of Things (RIoT) framework that integrates optical and radio wireless technologies with a focus on energy efficiency, scalability, and adaptability. To address the inherent complexity of hybrid optical-radio environments, a high-fidelity Digital Twin (DT) is developed within the Network Simulator 3 (NS-3) platform. The DT models deploy subsystems of the RIoT architecture, including Radio Frequency (RF) communication, Optical Wireless Communication (OWC), and energy harvesting and consumption mechanisms that enable autonomous operation. Real-time energy and power measurements from target hardware platforms are also incorporated to ensure accurate representation of physical behavior and enable runtime analysis and optimization. Building on this foundation, a proactive cross-layer optimization strategy is devised to balance energy efficiency and quality of service (QoS). The strategy dynamically reconfigures RIoT nodes by adapting transmission rates, wake/sleep scheduling, and access technology selection. Results demonstrate that the proposed framework, combining digital twin technology, hybrid optical-radio integration, and data-driven energy modeling, substantially enhances the performance, resilience, and sustainability of 6G IoT networks.

Abstract
A key feature of model finding techniques allows users to enumerate and explore alternative solutions. However, it is challenging to guarantee that the generated instances are relevant to the user, representing effectively different scenarios. This challenge is exacerbated in quantitative modelling, where one must consider both the qualitative, structural part of a model, and the quantitative data on top of it. This results in a search space of possibly infinite candidate solutions, often infinitesimally similar to one another. Thus, research on instance enumeration in qualitative model finding is not directly applicable to the quantitative context, which requires more sophisticated methods to navigate the solution space effectively. The main goal of this paper is to explore a generic approach for navigating quantitative solution spaces and showcase different iteration operations, aiming to generate instances that differ considerably from those previously seen and promote a larger coverage of the search space. Such operations are implemented in QAlloy - a quantitative extension to Alloy - on top of Max-SMT solvers, and are evaluated against several examples ranging, in particular, over the integer and fuzzy domains.

Abstract
Diabetic retinopathy (DR) is a leading cause of vision loss in working-age adults. While screening reduces the risk of blindness, traditional imaging is often costly and inaccessible. Artificial intelligence (AI) algorithms present a scalable diagnostic solution, but concerns regarding fairness and generalization persist. This work evaluates the fairness and performance of image-trained models in DR prediction, as well as the impact of disentanglement as a bias mitigation technique, using the diverse mBRSET fundus dataset. Three models, ConvNeXt V2, DINOv2, and Swin V2, were trained on macula images to predict DR and sensitive attributes (SAs) (e.g., age and gender/sex). Fairness was assessed between subgroups of SAs, and disentanglement was applied to reduce bias. All models achieved high DR prediction performance in diagnosing (up to 94% AUROC) and could reasonably predict age and gender/sex (91% and 77% AUROC, respectively). Fairness assessment suggests disparities, such as a 10% AUROC gap between age groups in DINOv2. Disentangling SAs from DR prediction had varying results, depending on the model selected. Disentanglement improved DINOv2 performance (2% AUROC gain), but led to performance drops in ConvNeXt V2 and Swin V2 (7% and 3%, respectively). These findings highlight the complexity of disentangling fine-grained features in fundus imaging and emphasize the importance of fairness in medical imaging AI to ensure equitable and reliable healthcare solutions.

Abstract
The rapid advancement of wireless technologies, particularly Wi-Fi, has spurred significant research into indoor human activity detection across various domains (e.g., healthcare, security, and industry). This work explores the non-invasive and cost-effective Wi-Fi paradigm and the application of deep learning for human activity recognition using Wi-Fi signals. Focusing on the challenges in machine interpretability, motivated by the increase in data availability and computational power, this paper uses explainable artificial intelligence to understand the inner workings of transformer-based deep neural networks designed to estimate human pose (i.e., human skeleton key points) from Wi-Fi channel state information. Using different strategies to assess the most relevant sub-carriers (i.e., rollout attention and masking attention) for the model predictions, we evaluate the performance of the model when it uses a given number of sub-carriers as input, selected randomly or by ascending (high-attention) or descending (low-attention) order. We concluded that the models trained with fewer (but relevant) sub-carriers are competitive with the baseline (trained with all sub-carriers) but better in terms of computational efficiency (i.e., processing more data per second).