Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Jordi Mateo has a total of 101 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Communication-Aware Collaborative Graph Neural Network Inference: Accuracy and Serialized Traffic Under Partition-Aware Emulation

    by Oriol Agost, Josep Rius, Jordi Vilaplana, Jordi Mateo

    Representation compression alone does not determine serialized traffic in partitioned graph neural network (GNN) inference. Deduplication, partition topology, node identifiers, and batch framing also affect the wire size. We measure these effects for a one-layer graph attention network that transmits task-trained embeddings. Validation selected 64-dimensional embeddings across five seeds on Cora, CiteSeer, PubMed, and Flickr. These embeddings retained 97.9%, 97.7%, and 99.2% of full-feature accuracy; Flickr achieved 102.4% of full-feature accuracy. Under four balanced-random partitions, their single-round communication reduction factors relative to a one-time centralized float32 upload were 12.69×, 41.30×, 5.19×, and 3.01×. The predefined criterion of 95% accuracy retention and a minimum fourfold single-round communication reduction factor therefore passed on three datasets. Int8 transmission reduced the payload by a factor of four, but reduced wire traffic by only 3.66–3.67× because identifiers and framing remained and each nonempty int8 batch added a 4-byte scale. A centralized int8-upload comparator preserved full-feature accuracy within 0.0006. Relative to this comparator, float32 collaboration achieved single-round communication reduction factors of 3.19×, 10.34×, 1.31×, and 0.76×, with 0.76× indicating increased traffic on Flickr. A post-review balanced METIS baseline reduced four-partition serialized traffic relative to random placement by 89.6%, 97.6%, 88.9%, and 44.7%, respectively, under three placement seeds. Constant-width depth extrapolation reduced the four-layer communication factors to 3.17×, 10.33×, 1.30×, and 0.75×; deeper-model accuracy was not measured. This study uses partition-aware emulation on one workstation and does not measure physical devices or energy. Its main contribution is a reproducible wire-accounting protocol that distinguishes representation size, payload size, serialized traffic, host cost, and modeled link delay.

  • Open Access

    A blockchain-based governance layer for verifiable bid integrity in electricity market clearing

    by Ferran Aran, Pablo Fraile, Sambeet Mishra, Josep Rius, Jordi Mateo

    Electricity markets depend on centralized clearing mechanisms that require participants to trust that submitted bids are preserved and accurately incorporated into the market-clearing process. Current blockchain-based energy market solutions either decentralize the auction mechanism or utilize the blockchain solely as a transaction log, lacking verifiable assurances that off-chain clearing employs the complete and unaltered set of submitted bids. This work introduces a hybrid blockchain-based governance architecture that enables verifiable bid integrity for centralized electricity market clearing while maintaining conventional off-chain clearing procedures. The architecture records cryptographic commitments of submitted orders on a permissioned Hyperledger Fabric blockchain, stores clear-text bids in restricted private collections, and anchors settlement outputs on-chain via an oracle interface. This design allows independent post-clearing verification that the orders used in clearing correspond precisely to those committed before auction closure, without disclosing confidential bid information. The system is evaluated using a real intraday electricity market dataset containing 46,643 orders from a full trading day in the Spanish market. Experimental results show that the architecture maintains one-to-one correspondence between submitted orders and on-chain commitments, enforces correct market lifecycle transitions, and detects inconsistencies between committed bids and the inputs used during clearing, providing tamper-evident guarantees of input integrity in adversarial scenarios. Performance benchmarking against a centralized database baseline shows that the blockchain implementation introduces additional latency and achieves 176.5 transactions per second under the evaluated configuration, reflecting a throughput-limited regime while remaining compatible with realistic intraday auction time windows. These findings demonstrate that blockchain technology can serve as a practical governance layer for electricity markets by shifting trust from unverifiable operator actions to cryptographically auditable input integrity, without requiring modifications to existing clearing algorithms. The approach does not verify the correctness of the clearing algorithm itself but ensures that its inputs are cryptographically auditable.

  • Open Access

    The Trustworthy Model Context Protocol (MCP) Registry: An Architectural Blueprint for Cryptographic Provenance and Runtime Integrity

    by Lluis Mas, Jordi Vilaplana, Josep Rius, Radu Spaimoc, Jordi Mateo

    The Model Context Protocol (MCP) enables Large Language Models (LLMs) to act as autonomous agents that orchestrate complex workflows over distributed systems, while MCP resolves integration bottlenecks by standardizing agent-to-resource communication. Its current registry relies on an unverified pointer architecture, exposing agentic workflows to supply chain poisoning and dynamic capability mutation (“Rug Pull”) attacks. This paper identifies this gap and proposes a three-layer architectural framework for a Trustworthy MCP Registry. The novelty of our contribution lies not in the individual standards employed (RFC 8615, Sigstore, and JCS/JWS are established technologies), but in their specific composition to address MCP’s unique runtime security requirements: (1) RFC 8615 Well-Known URIs for decentralized server discovery and domain-bound identity; (2) Sigstore Keyless signing to bind server artifacts to audited CI/CD environments without managing long-lived keys; and (3) JSON Canonicalization Scheme (RFC 8785) combined with JWS to provide deterministic, per-message integrity verification of live capability updates. We present a prototype implementation and an experimental evaluation conducted in a controlled, synthetic environment. Results indicate that the cryptographic overhead averages 0.61 ms per signing operation and that the Layer 3 mechanism correctly rejects all 100 simulated Rug Pull attempts, as expected by construction, since an attacker without the server’s private key cannot produce a valid signature. These findings suggest that the proposed approach is feasible and warrants further evaluation in real-world deployment settings.


  • Open Access

    OptiScale: A data-driven framework for selecting spatial and temporal resolution in biodiversity analysis

    by Rui Alves, Emilce Soledad Martinez, Eva Tejada-Gutiérrez, Abderrhamane Eleiwa, Albert Sorribas, Jordi Mateo Fornes, Francesc Solsona, Martin Calcaterra, Alberto Marin-Sanguino, Jesús Asín, et al.

    Biodiversity analysis based on species-occurrence data is highly sensitive to the spatial and temporal resolution at which data are aggregated, leading to scale-dependent biases associated with the Modifiable Area and Temporal Unit Problems (MATU/MTUP). Despite their importance, resolution choices are often arbitrary, potentially introducing artefactual relationships in the data and limiting the reproducibility and comparability of ecological studies. Here, we present OptiScale, a data-driven framework designed to systematically identify optimal spatial-temporal resolution (OpR) for biodiversity analysis. The workflow integrates species-area relationship (SAR) and piecewise regression analysis to detect candidate spatial resolution, and process control analysis to evaluate temporal stability. Then, a multicriteria selection procedure is applied to identify the OpR that balances data coverage, fragmentation, and statistical robustness. OptiScale was evaluated across three case studies with contrasting ecological and sampling context. Results demonstrate that OpR is context-dependent and emerges from trade-offs between spatial aggregation, temporal stability, and data sparsity. Overall, OptiScale reduces aggregation bias by replacing arbitrary resolution choices with an explicit and statistically grounded procedure. The framework is scalable and adaptable to different datasets. Beyond plant biodiversity, its principles are transferable to any georeferenced dataset, which helps make large-scale ecological analyses more robust and comparable.

  • Open Access

    Electrolyte Imbalance and Indirect Indicators of Dehydration in Temporary Agricultural Workers Exposed to Extreme Heat in the Mediterranean: An Observational Study on Environmental Health Risks

    by Tania Cemeli, Glòria Tort-Nasarre, Judith Roca, Ana Lavedán-Santamaría, Carme Campoy, Laia Selva-Pareja, Jordi Vilaplana, Jordi Mateo, Anna Espart

    Background: Climate change is intensifying extreme heat exposure in Mediterranean agricultural systems. Migrant workers engaged in outdoor fieldwork are a highly vulnerable population with limited access to resources. Crucially, there is a notable lack of data on how heat affects these workers in this specific region. Objective: This study aimed to analyze the physiological effects of high-temperature exposure by quantifying and correlating indirect indicators of dehydration and electrolyte imbalance (sodium and potassium losses, sweat, body weight, and blood pressure). Methods: An observational study was conducted over nine consecutive days involving ten agricultural participants, yielding 90 observations. Measurements of body weight, heart rate, blood pressure, skin temperature, sweat loss, and sodium and potassium concentrations were taken before, during, and after daily field activity. Results: Results showed considerable interindividual variability in thermophysiological responses. Participants lost an average of 0.8 kg (range –9.1 to +3.6 kg) and produced 3.91 L of sweat (range 1.9–6.4 L), with sodium and potassium losses of 4932 mg and 646 mg, respectively. Sweat loss correlated with sodium (r = 0.414, p = 0.001) and potassium (r = 0.791, p < 0.001), and diastolic blood pressure was moderately associated with weight loss (r = 0.576, p = 0.016). Conclusions: Sweat loss was the main driver of electrolyte depletion, with marked interindividual variability. Monitoring sweat-related indicators and diastolic blood pressure could help detect dehydration risk in agricultural workers exposed to extreme heat. Targeted hydration strategies and occupational health education are essential to mitigate these risks.


  • Open Access

    Early detection of vascular catheter-associated infections employing supervised machine learning - a case study in Lleida region

    by Radu Spaimoc, Jordi Mateo, Francesc Solsona, Alfredo Jover-Sáenz, Fernando Barcenilla, María Ramírez-Hidalgo, Marcos Serrano, Miquel Mesas, Dídac Florensa

    Healthcare-associated infections (HAIs), particularly Vascular Catheter-Associated Infections (VCAIs), are a significant concern, accounting for over 7% of all infections and are often linked to medical devices. Early detection of VCAIs before invasive infection is crucial for improving hospital care and reducing antibiotic use. This study retrospectively developed and evaluated machine learning models to classify VCAIs from patient medical records, excluding fever and antibiotic prescription indicators. The dataset, collected from the group of public hospitals of the Lleida health region in Catalonia (Spain) between 2011 and 2019, consisted of 24,239 episodes with 150 features related to vascular catheter use. After validation, processing and feature engineering, the dataset showed an imbalance, with 94.46% (10,090) non-catheter episodes and 5.53% (591) catheter infection cases. Machine learning classifiers demonstrated significant challenges in classifying imbalanced datasets, particularly in the context of VCAIs. While most models achieved high accuracy and specificity (approximately 97%), they frequently exhibited limited sensitivity, reaching only around 60% in the best-performing cases. Among the evaluated classifiers, the Gradient Boosting (GB) model outperformed others, attaining the highest balanced accuracy (82.5%) and sensitivity (67%), underscoring its potential utility for early VCAI detection. Additionally, the analysis examined the impact of oversampling techniques on model performance. Although these methods enhanced metrics for some classifiers, they did not consistently outperform models trained on the original dataset. Therefore, if the improvement is not significant, it is preferable to use the original dataset. This study highlights that strategic feature engineering with the GB classifier is sufficient to obtain robust VCAI detection before the appearance of a probable sepsis.


  • Open Access

    Early detection of Vascular Catheter-Associated Infections employing supervised machine learning - A case study in Lleida Region

    by Radu Spaimoc, Jordi Mateo, Francesc Solsona, Alfredo Jover-Sáez, Fernando Barcenilla, María Ramírez, Marcos Serrano, Miquel Mesas, Dídac Florensa

    Healthcare-associated infections (HAIs), particularly Vascular Catheter-Associated Infections (VCAIs), are a significant concern, accounting for over 7% of all infections and are often linked to medical devices. Early detection of VCAIs before invasive infection is crucial for improving hospital care and reducing antibiotic use. This study retrospectively developed reliable machine learning models to clasify VCAIs from patient medical records, excluding fever and antibiotic prescription indicators. The dataset, collected from the group of public hospitals of the Lleida health region in Catalonia (Spain) between 2011 and 2019, consisted of 24,239 episodes with 150 features related to vascular catheter use. After validation, processing and feature engineering, the dataset showed an imbalance, with 94.46% (10,090) non-catheter episodes and 5.53% (591) catheter infection cases. The study’s results underscore that classifiers have demonstrated respectable performance postpreprocessing and feature engineering despite the initial imbalance within the dataset, with balanced accuracies ranging between 70% and 80%. Notably, the Decision Tree (DT) classifier emerged as the frontrunner, achieving 82%, thereby validating its effectiveness. While various oversampling strategies, such as SMOTE, were explored, these did not significantly strengthen the performance beyond what was achieved using DT alone. This study highlights that strategic feature engineering with the DT classifier is sufficient to obtain robust VCAI detection before the apparition of a probable sepsis.

  • Predicting the Colorectal Cancer Mortality in the Region of Lleida, Spain: A Machine Learning Study

    by Didac Florensa, Jordi Mateo, Francesc Solsona, Pere Godoy, Leonardo Espinosa-Leal

  • Open Access

    Resilient edge predictive analytics by enhancing local models

    by Qiyuan Wang, Saleh Alfahad, Jordi Mateo Fornes, Christos Anagnostopoulos, Kostas Kolomvatsos

    In distributed computing environments, the collaboration of nodes for predictive analytics at the network edge plays a crucial role in supporting real-time services. When a node’s service becomes unavailable for various reasons (e.g., service updates, node maintenance, or even node failure), the rest of the available nodes connot efficiently replace its service due to different data and predictive models (e.g., machine learning [ML] models). To address this, we propose decision-making strategies rooted in the statistical signatures of nodes’ data. Specifically, these signatures refer to the unique patterns and behaviors within each node’s data that can be leveraged to predict the suitability of potential surrogate nodes. Recognizing and acting on these statistical nuances ensures a more targeted and efficient response to node failures. Such strategies aim to identify surrogate nodes capable of substituting for failing nodes’ services by building enhanced predictive models. Our resilient framework helps to guide the task requests from failing nodes to the most appropriate surrogate nodes. In this case, the surrogate nodes can use their enhanced models, which can produce equivalent and satisfactory results for the requested tasks. We provide experimental evaluations and comparative assessments with baseline approaches over real datasets. Our results showcase the capability of our framework to maintain the overall performance of predictive analytics under nodes’ failures in edge computing environments.

  • Open Access

    Exploring Cancer Incidence, Risk Factors, and Mortality in the Lleida Region: Interactive, Open-source R Shiny Application for Cancer Data Analysis

    by Didac Florensa, Jordi Mateo-Fornes, Sergi Lopez Sorribes, Anna Torres Tuca, Francesc Solsona, Pere Godoy

    Background: The cancer incidence rate is essential to public health surveillance. The analysis of this information allows authorities to know the cancer situation in their regions, especially to determine cancer patterns, monitor cancer trends, and help prioritize the allocation of health resource. Objective: This study aimed to present the design and implementation of an R Shiny application to assist cancer registries conduct rapid descriptive and predictive analytics in a user-friendly, intuitive, portable, and scalable way. Moreover, we wanted to describe the design and implementation road map to inspire other population registries to exploit their data sets and develop similar tools and models. Methods: The first step was to consolidate the data into the population registry cancer database. These data were cross validated by ASEDAT software, checked later, and reviewed by experts. Next, we developed an online tool to visualize the data and generate reports to assist decision-making under the R Shiny framework. Currently, the application can generate descriptive analytics using population variables, such as age, sex, and cancer type; cancer incidence in region-level geographical heat maps; line plots to visualize temporal trends; and typical risk factor plots. The application also showed descriptive plots about cancer mortality in the Lleida region. This web platform was built as a microservices cloud platform. The web back end consists of an application programming interface and a database, which NodeJS and MongoDB have implemented. All these parts were encapsulated and deployed by Docker and Docker Compose. Results: The results provide a successful case study in which the tool was applied to the cancer registry of the Lleida region. The study illustrates how researchers and cancer registries can use the application to analyze cancer databases. Furthermore, the results highlight the analytics related to risk factors, second tumors, and cancer mortality. The application shows the incidence and evolution of each cancer during a specific period for gender, age groups, and cancer location, among other functionalities. The risk factors view permitted us to detect that approximately 60% of cancer patients were diagnosed with excess weight at diagnosis. Regarding mortality, the application showed that lung cancer registered the highest number of deaths for both genders. Breast cancer was the lethal cancer in women. Finally, a customization guide was included as a result of this implementation to deploy the architecture presented. Conclusions: This paper aimed to document a successful methodology for exploiting the data in population cancer registries and propose guidelines for other similar records to develop similar tools. We intend to inspire other entities to build an application that can help decision-making and make data more accessible and transparent for the community of users.