Conference Agenda

This work presents a secure cloud computing application that leverages homomorphic encryption to enable encrypted data processing without compromising confidentiality. The application consists of a desktop client and a RESTful backend API; The API is designed to support statistical analysis and machine learning operations on encrypted datasets. The client handles key generation, model training on plaintext data, and encryption of data using BFV or CKKS schemes. Non-numerical data is encrypted using AES for efficiency. The server performs fully homomorphic computations on encrypted data without ever accessing plaintext values. The proposed application demonstrates the advantages of homomorphic encryption over traditional encryption methods, particularly can be used in privacy-sensitive sectors such as healthcare, finance, and government (e.g. electronic voting). Using the TenSEAL library, this work validates the feasibility of secure inference in machine learning applications, despite current limitations such as lack of support for operations like square root and logical comparisons. The application is extensible and suitable for integration with databases and third-party applications and can be used as a reusable SDK for future works.

Liver cirrhosis is a leading cause of morbidity and mortality worldwide, with an increasing prevalence associated with multiple etiologies. Accurate prediction of survival in cirrhotic patients is crucial for risk stratification and the optimization of therapeutic resources, particularly in identifying candidates for liver transplantation. This study comparatively evaluated three machine learning approaches: Random Forest with class weights, Artificial Neural Network with SMOTE oversampling, and a Fuzzy Logic classifier with reinforced rules, using the public dataset from the Mayo Clinic Trial (n=8,181). The class distribution showed extreme imbalance: 62.5% censored, 33.9% deceased, and 3.6% transplanted. The results showed that Random Forest achieved the best overall performance (Balanced Accuracy=0.652, F1-macro=0.666), with particularly outstanding accuracy in the majority classes (F1-censored=0.866, F1-death=0.760).

This systematic literature review examines the advances and challenges of Generative Artificial Intelligence (GenAI) in engineering using the PICOC strategy and strictly following PRISMA 2020 guidelines to ensure transparency, reproducibility, and methodological rigor. The search conducted in Scopus (2023–2026) initially identified 24,166 records, which were filtered through predefined inclusion and exclusion criteria, resulting in a final sample of 98 studies. Findings indicate that GenAI applications are predominantly concentrated in industrial and chemical process optimization (59.7%), followed by generative design and simulation (19.4%), predictive maintenance (6.2%), and telecommunications (5.4%). Heuristic optimizers dominate methodological approaches (90.7%), complemented by advanced generative models (76.7%), Deep Learning (34.1%), and traditional Machine Learning (27.9%). Only 60.5% of the studies report experimental comparisons, revealing gaps in methodological standardization. Reported performance improvements include efficiency gains (54.3%), productivity increases (24%), and enhanced reliability (14%). Validation practices rely primarily on statistical cross-validation (91.5%) and real-world industrial testing (59.7%). However, persistent limitations involve data scarcity (23.3%) and high computational costs (14%). Overall, the evidence confirms that GenAI is emerging as a strategic technology for addressing complex engineering problems, although broader industrial validation frameworks and standardized evaluation metrics remain necessary to ensure scalable and sustainable adoption.

Defect detection on metallic surfaces using artificial vision systems is highly sensitive to environmental disturbances such as reflections, shadows, and color temperature variations. These factors directly affect the performance of artificial neural networks (ANNs) during training and inference. This study evaluates the impact of three illumination ranges—optimal (150–300 Lux), moderate (350–500 Lux), and inadequate (550–1200 Lux)—on the performance of YOLO V5, YOLO V8, and Roboflow 3.0 for metallic surface defect detection. A dataset was collected from local metalworking workshops, and a spiral methodology was implemented to iteratively evaluate model behavior. Results show that optimal illumination conditions significantly improve performance, reaching a maximum mAP of 99.5% for iron surfaces and 88.7% for aluminum. The findings demonstrate that illumination control plays a critical role in ensuring reliable defect detection, particularly for highly reflective materials.

The growing adoption of electronic invoicing has brought operational benefits but also increased cybersecurity risks. This study conducts a systematic literature review of scientific publications from 2020 to 2025 to identify cyberattack mitigation strategies applied to electronic invoicing systems. The PRISMA approach and the PICO model were used to structure the search, resulting in the selection of 62 relevant articles from Scopus and Dimensions. The findings reveal a predominance of preventive approaches based on artificial intelligence, blockchain, and big data. Protection gaps were also identified in sectors such as SMEs and retail. The study proposes future research directions focused on integrating predictive tools with adaptive regulatory frameworks.

This paper examines the transition from traditional, reactive cybersecurity models toward a strategic paradigm in which security functions as a value-generating organizational asset within the digital economy. To operationalize this shift, it introduces the Decentralized, Transparent, and Tokenized (DTT) Strategic Security Framework, an original conceptual model that integrates blockchain architectures, verifiable transparency mechanisms, and tokenization processes to redefine digital protection. The framework articulates how decentralization enhances systemic resilience by reducing single points of failure, transparency strengthens accountability and regulatory trust through immutable auditability, and tokenization enables the economic activation of security-related artifacts such as credentials, access rights, and verifiable proofs. Drawing on illustrative international and Argentine case studies across high-value sectors—including finance, energy, logistics, agroindustry, technology, and Small and Medium-Sized Enterprises—the paper demonstrates how early adoption of DTT principles improves operational integrity, accelerates compliance, and supports new trust-based business models. The main contribution of this work is the presentation of a structured and replicable framework that bridges technical feasibility with strategic and economic rationale, offering organizations a pathway to transform security from a defensive cost center into a source of tangible and intangible value. The findings suggest that DTT-aligned security architectures strengthen resilience, enhance market differentiation, and support sustainable digital transformation, particularly in emerging economy contexts.