Conference Agenda

Advances in Explainable Artificial Intelligence (XAI) have transformed the field of medical diagnosis, addressing challenges related to interpretability and trust in deep learning models. This study conducts a systematic literature review (SLR) to explore how XAI has been applied in the analysis of radiological images, such as X-rays, computed tomography, and magnetic resonance imaging, with the goal of providing transparency in diagnostic outcomes. The analysis identified a predominance of post-hoc techniques, such as SHAP and LIME, alongside model-inherent approaches, including decision trees and neural networks with attention mechanisms. These tools have enhanced trust in AI systems by offering clear interpretations of algorithmic reasoning. However, significant gaps were identified in the standardization of evaluation metrics and the suitability of explanations for clinical professionals. XAI represents a critical step toward the widespread acceptance of artificial intelligence in medical diagnosis, addressing the challenges of deep learning model opacity, consolidating existing initiatives, and offering key recommendations for future research, with an emphasis on developing standardized metrics and explainability tools focused on the needs of medical professionals and their patients.

Major depression represents a critical challenge in public health due to its high prevalence and impact on patients’ quality of life. This study proposes a predictive model based on Gradient Boosting Trees (GBT) to prioritize clinical care for patients with potential major depression in Lima, Peru. A dataset with 10,000 clinical records was used, including demographic, behavioral, and medical variables, such as sleep patterns, family history, and lifestyle habits. The applied methodology comprises data preprocessing, feature selection, hyperparameter optimization, and validation using metrics such as AUC-ROC, accuracy, and F1-score. The model obtained an accuracy of 89.7%, an AUC-ROC of 0.92, and a 30% reduction in diagnostic time compared to traditional methods. In addition, it allowed a 35% improvement in the identification of high-risk patients, optimizing the allocation of medical resources. These results demonstrate that the use of machine learning in mental health can significantly improve the efficiency of depression detection and treatment. The implementation of this model in hospitals and public health centers is recommended to strengthen clinical decision-making and ensure more equitable and evidencebased care.

This study conducts a systematic review of the implementation of business intelligence (BI) systems in the retail sector, with the aim of identifying key tools and methodologies to optimize customer retention. The PRISMA methodology was applied, analyzing 25 articles selected from a total of 190 initial studies obtained in Scopus, ScienceDirect, IEEE Xplore, Google Scholar and ResearchGate. The findings reveal that, in international markets, the adoption of Big Data, Cloud Computing and methodologies such as CRISP-DM allows for personalizing the customer experience, improving retention and optimizing decision making. However, in developing countries, the adoption of BI is more limited, with accessible tools such as Power BI and SQL Server predominating, which restricts the use of predictive models and the management of large volumes of data. Significant gaps are identified in infrastructure and application of advanced BI models, which limits the competitiveness of local companies in a data-driven global market. Recommendations include promoting future research on the integration of advanced technologies in SMEs and strategies to strengthen their competitiveness in digital environments. The results underline the importance of adapting BI solutions to local needs to improve loyalty and personalization in retail.

Currently, technological development through digital tools is transforming the activities of legal operators in judicial processes, especially in the criminal process; in this sense, the objective of this research is “to determine how digital tools are being applied in the criminal process” considering the advance of ICT. The methodology used is qualitative, descriptive and exploratory approach through the search and review of primary and secondary scientific articles of high impact indexed in the databases Scopus, Scielo, Dialnet and Redalyc; for this purpose the Boolean operator AND has been used, using as keywords “criminal process”, “digital tools”, “digital tools and criminal process”, “digital tools and justice administration”, “justice and digital” and “digital tools and justice”; English and Spanish language, with a period of time from 2018 to 2024. From the results collected and analyzed, it was evidenced that digital tools support and facilitate the activity of justice operators in criminal law and the responsibility of the State in developing legal norms that regulate the creation of algorithms for this purpose, with the aim of protecting fundamental rights of users.

In this article, a future comprehensive solution for vehicular traffic management was implemented, focusing on the automation of vehicle identification and classification. To do this, a camera was configured with ANPR (Automatic License Plate Recognition) technology that captures images of the license plates and extracts the alphanumeric information. This information, along with the images, is transmitted through a system designed with Amazon Web Services (AWS) cloud services, such as Kinesis Video Streams, EC2 and S3, which allow efficient transmission, processing and storage of the data. In addition, OpenCV opensource software was used to establish the connection between the camera and the system, which contributed to the cost optimization of the project. The agile SCRUM methodology and the Logical Framework were fundamental for the management of the project, allowing a clear identification of the central problem and efficient execution of the process. As a result of this implementation, a significant improvement of 80% was obtained in the time needed to identify and classify vehicles, which translates into more agile and efficient traffic management

The Internet has the problem known as the spread of fake news. Based on this, the authors have chosen the social network Twitter to be studied because it is known as the medium in which false news is frequently disseminated. Therefore, this investigation elaborates on a data set using natural language processing. It comprises 1600 tweets classified as true or false according to their content and based on news verification articles from Perú. With this data set, four classification models are designed with high precision to identify if a tweet is true or false, using first Natural Language Processing, Logistic Regression, Support Vector Machine, Dense Neural Network, and Random Forests algorithms. Then, the hyperparameters of all algorithms are tuned. Finally, after the performance evaluation of the classification models, the authors recommend the Support Vector Machine as the best algorithm.