Editorial

Abstract

Latin-American Journal of Computing (LAJC), Vol. XIII, Issue 2, July - December 2026

It is a true privilege to welcome our readers to this new issue of the Latin-American Journal of Computing (LAJC) (Volume XIII, Issue 2). With this edition, we reaffirm our commitment to serving as a key meeting point for debate and the dissemination of serious, rigorous research with a real impact on the scientific community. On this occasion, we gather eight original articles that showcase where computer science is heading today, striking a balance between theory and practical solutions for highly relevant topics such as artificial intelligence, cybersecurity, environmental care, and the digital transformation of our cities.

What makes the works in this volume special is that they do not just stay on paper; they all seek to connect methodological rigor with real-world applications. For instance, the field of natural language processing and generative AI brings us two valuable proposals: the first introduces the PAGE (Prompt Augmentation for text Generation Enhancement) framework, an ingenious alternative that improves control and quality in Large Language Models (LLMs) without having to commit the massive resources required by traditional fine-tuning. The second proposal offers a deep analysis of how machine learning classifiers behave when performing sentiment analysis on Spanish tweets, focusing on the resilience of these models when faced with the challenge of working with highly imbalanced data in corporate contexts.

Cybersecurity also takes center stage through two well-founded approaches. The first breaks traditional defense paradigms by proposing a high-interaction Honeypot system that uses the GPT-4o model to deceive attackers in real time, mimicking a Linux terminal and classifying incoming commands as safe, suspicious, or malicious to retain the intruder as long as possible without risking the network. Complementing infrastructure security, the second article outlines a lightweight intrusion detection system based on optimized Random Forests, specifically designed to provide accurate and easy-to-interpret responses in industrial environments that lack extensive hardware resources.

On the other hand, the combination of intelligent systems and computer vision delivers direct solutions for health and the environment. Standing out in this space is a framework that couples the Internet of Things (IoT) with Deep Reinforcement Learning (DRL) under the Proximal Policy Optimization (PPO) strategy, successfully automating and prioritizing the flow and management of hazardous hospital waste. Added to this is a systematic literature mapping focused on public health, which explores the use of Convolutional Neural Networks (CNNs) to automate the morphological classification of hematophagous diptera, demonstrating the massive potential of computer vision in supporting digital taxonomy.

The closing of this edition brings us toward mathematical optimization and the analysis of technology's social impact. On the theoretical side, the PBI-BFS-MaOA multi-objective evolutionary algorithm is introduced—a proposal that implements a local selection rule on the boundary front to keep selection pressure under control when dealing with high-dimensional problems. Bringing this down to the reality of our societies, a detailed documentary study hits the nail on the head by analyzing the technical challenges, lack of digital literacy, and barriers to exclusion that hinder the use of mobile applications intended for mental health and public safety within the urban transformation plans of smart cities in our region.

Looking at the final work of this volume reminds us that scientific research truly makes sense when it responds to the social and human needs of our surroundings. We would like to thank the authors for choosing our journal to publish their findings, the reviewers for their enormous and meticulous peer-review work, and you, our readers, for continuing to drive the growth of science across different fields of knowledge through your reading.

Dr. Marco Sánchez

Latin-American Journal of Computing (LAJC)

Escuela Politécnica Nacional, Ecuador