Martinez-Villaseñor, Lourdes
Main Affiliation
Preferred name
Martinez-Villaseñor, Lourdes
Official Name
Martínez Villaseñor, María de Lourdes
ORCID
0000-0002-9038-7821
Researcher ID
N-7607-2018
Scopus Author ID
55521085100
94 results
Now showing 1 - 10 of 94
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Ethical Challenges in Demand Prediction: A Case Study in the Wholesale Grocery Sector(Instituto Politécnico Nacional. Centro de Investigación en Computación, 2025) - Some of the metrics are blocked by yourconsent settings
Item type:Publication, - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Human-Friendly Explanation Model Based on the Aristotelian Practical Syllogism for Reinforcement Learning Agents in Urban Intelligent Design(Springer Nature Switzerland, 2025); Crespo Murillo, AlanThis paper introduces a novel explanation model for reinforcement learning agents, grounded in Aristotle’s practical syllogism, with the goal of enhancing human-friendly explanations (HFEs) in urban intelligent design. By aligning the internal processes of a Q-learning agent with Aristotelian categories such as telos, phronēsis, and boulesis, the model enables the interpretation of agent behavior as a form of ethically structured practical reasoning. The selected case study focuses on an urban planning agent designed to optimize service accessibility and spatial coherence in real city environments. Through this alignment, the agent’s policy learning, state-action evaluation, and reward optimization are made intelligible to human users in terms of goals, normative principles, perception, deliberation, and action. Evaluated using the Human-Friendly Explanations (HFE) checklist, the model exhibits strengths in interpretability, comprehensibility, and ethical relevance, while identifying areas for improvement such as contrastive reasoning, personalization, and regulatory compliance. This work offers a conceptual and methodological foundation for integrating philosophical models into Explainable Reinforcement Learning (XRL), facilitating transparent, ethical, and user-aligned AI systems. Future directions include empirical validation, interactive implementation, and domain-specific adaptation across sociotechnical contexts. ©The authors ©Springer. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Risk Factors for Hypertension and Health Policy(Springer Nature Switzerland, 2025); Salazar Altamirano, VladimirHypertension (HTN) remains a leading cause of global morbidity and mortality, with its prevalence rapidly increasing in middle-income countries like Mexico. This study leverages data from the National Health and Nutrition Survey (ENSANUT) 2022–2023 to identify significant risk factors for HTN using machine learning models and explainable artificial intelligence techniques. Methods: We analyzed 8,650 Mexican adults aged 20 years using logistic regression, random forest, and extreme gradient boosting (XGBoost) classifiers, optimized through Optuna framework. Hypertension was defined using ACC/AHA 2017 criteria (130/80 mmHg or previous diagnosis). SHAP, LIME, and ELI5 were employed for model interpretation. Performance was evaluated using stratified 5-fold cross-validation and external validation on ENSANUT 2023 data. Results: XGBoost demonstrated superior predictive performance (ROC-AUC: 0.758, F1-score: 0.703) compared to logistic regression and random forest. Key risk factors identified were age, body mass index (BMI), gender, family history of hypertension (maternal and paternal), and physical activity—particularly sedentary behavior. SHAP-based subgroup analysis revealed that age and BMI consistently emerged as the most influential factors across both genders, with family history of hypertension (maternal and paternal) also showing significant importance. Gender-specific differences were subtle, with similar risk factor patterns observed in men and women. Conclusions: Machine learning models, particularly XGBoost with explainable AI techniques, effectively predict hypertension risk using nationally representative Mexican data. These findings support implementation of ML-driven risk stratification in primary care and inform targeted public health interventions. The integration of XAI methods provides transparent, interpretable predictions suitable for clinical decision-making and policy development in Mexico’s healthcare system. ©The authors ©Springer. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, A Novel Ethical Design Framework Applied to Image Classification Challenges in the Fashion Industry(Springer Nature Switzerland, 2025); As artificial intelligence (AI) continues to play a pivotal role in image classification applications, the ethical implications of these technologies become increasingly significant. This paper explores the intersection of AI and ethics in the context of image classification, specifically focusing on the application of ethical design principles through a framework for a use of case in the fashion industry involving bags images and social media. This work delves into the integration of a comprehensive ethical framework around all the design process. The case study involves the development and implementation of a neural network tailored for bag image classification, leveraging transfer learning techniques. Through a meticulous examination of the ethical dimensions inherent in image classification, the study aims to establish a foundation for responsible and transparent AI practices. ©The authors ©Springer. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Preface : Artificial Intelligence – COMIA 2025 17th Mexican Congress, Mexico City, Mexico, May 12–16, 2025, Proceedings, Part I(Springer Science and Business Media Deutschland GmbH, 2025); - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Contrastive Steering Vectors for Autoencoder ExplainabilityGenerative models, particularly autoencoders, often function as black boxes, making it challenging for non-expert users to effectively control the generation process and understand how inputs affect outputs. Existing methods for improving interpretability and control frequently require specific model training regimes or labeled data, limiting their applicability. This work introduces a novel approach to enhance the controllability and explainability of generative models, specifically tested on autoencoders with entangled latent spaces. We propose using a semi-supervised contrastive learning setup to learn steering vectors. These vectors, when added to an input’s latent representation, effectively manipulate specific attributes in the generated output without conditional training of the model or attribute classifiers, thus being applicable to pretrained models and avoiding compound classification errors. Furthermore, we leverage these learned steering vectors to interpret and explain the decoding process of a target attribute, allowing for efficient exploration of feature dimension interactions and the construction of an interpretable plot of the generative process, while lowering scalability limitations of perturbation-based Explainable AI (XAI) methods by reducing the search space. Our method provides an efficient pathway to controllable generation, offers an interpretable result of the model’s internal mechanisms, and relates the interpretations to human-understandable explanation questions. ©The authors ©MDPI AG ©Electronics. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, 7 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Process of Concept Alignment for Interoperability between Heterogeneous Sources(2013); González-Mendoza, MiguelSome researchers in the community of user modeling envision the need to share and reuse information scattered over different user models of heterogeneous sources. In a multi-application environment each application and service must repeat the effort of building a user model to obtain just a narrow understanding of the user. Sharing and reusing information between models can prevent the user from repeated configurations, help deal with application and services’ “cold start” problem, and provide enrichment to user models to obtain a better understanding of the user. But gathering distributed user information from heterogeneous sources to achieve user models interoperability implies handling syntactic and semantic heterogeneity. In this paper, we present a process of concept alignment to automatically determine semantic mapping relations that enable the interoperability between heterogeneous profile suppliers and consumers, given the mediation of a central ubiquitous user model. We show that the process of concept alignment for interoperability based in a two-tier matching strategy can allow the interoperability between social networking applications, FOAF, Personal Health Records (PHR) and personal devices.Scopus© Citations 2 4 2 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, 4