Lorenzo Jovanny Cevallos-Torres

Prof. Lorenzo Jovanny Cevallos-Torres


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 July 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
Maikel Leyva VázquezMaikel Leyva Vázquez
Maikel Leyva Vázquez
Dayron Rumbaut-RangelDayron Rumbaut-Rangel
Dayron Rumbaut-Rangel
LCLorenzo Jovanny Cevallos-Torres
Lorenzo Jovanny Cevallos-Torres
AMAlexis Matheu
Alexis Matheu
Florentin SmarandacheFlorentin Smarandache
Florentin Smarandache
Contemporary alignment methods—RLHF, Constitutional AI, Direct Preference Optimization—share an implicit epistemological commitment: that diverse human values can be aggregated into a single utility function or constitution fixed at training time. When principles are elicited from genuinely pluralistic populations, structurally irreducible contradictions emerge; current pipelines resolve them by consensus filters that function, in effect, as majority vote. We argue that this alignment collapse is not an implementation detail but a representational limit: standard uncertainty formalisms (Bayesian, fuzzy, Dempster–Shafer, imprecise probability), in their dominant normalized forms, do not preserve well-supported contradictory evidence without collapsing it into a single ordering. We propose a paraconsistent alternative in which reward signals, policy outputs, and constitutional evaluations are encoded as neutrosophic triplets (T, I, F) ∈ [0, 1]3 with no sum constraint, so that T + F > 1 becomes a representable—and auditable—state of genuine value conflict. We reformulate reward modeling and constitutional aggregation in this framework, introduce a deliberation-threshold decision functional, and report preliminary empirical evidence: across 2,320 evaluations of six frontier model families, models violate the classical sum constraint (T + I + F > 1) on 62–100% of contested baseline prompts and never on tautological controls, while the strict paraconsistent signature (T + F > 1) concentrates selectively on ethical-conflict items (92%) and expands under audit protocols (71%); which conflicts become visible depends strongly on which model adjudicates them (paired Cochran’s Q, p ≈ 10−12). We specify a two-arm training experiment that would confirm or refute the framework’s central operational claim, and close with four testable hypotheses and a research agenda that treats the question aligning with whom? as political and epistemic before it is technical.

20 October 2026 in Engineering Proceedings
Verified
Conference Paper
Open access
PDF file
LVLeonel Vasquez-Cevallos
Leonel Vasquez-Cevallos
LCLorenzo Jovanny Cevallos-Torres
Lorenzo Jovanny Cevallos-Torres
RCRosangela Caicedo-Quiroz
Rosangela Caicedo-Quiroz
RTRoberto Tolozano-Benites
Roberto Tolozano-Benites
FBFranklin Parrales Bravo
Franklin Parrales Bravo
Chronic kidney disease (CKD) necessitates ongoing dietary self-management; nevertheless, numerous patients have challenges in comprehending and implementing nutritional guidelines in their daily routines. This research introduces a mobile tutoring prototype utilizing a hybrid computational intelligence framework for dietary recommendations and health literacy assistance in CKD. The suggested methodology adheres to a Knowledge Discovery in Databases pipeline that includes data selection, preprocessing, transformation, data mining, evaluation, and software integration. The predictive core integrates a long short-term memory (LSTM) model with a neutrosophic adjustment layer to simultaneously tackle temporal dynamics and uncertainty in patient-related data. The system employs clinically pertinent nutritional and health metrics, encompassing salt and potassium consumption, body mass, hydration status, blood pressure, creatinine levels, glomerular filtration rate, and urinary protein. In the initial assessment, the hybrid model surpassed independent methods, attaining 92% precision, whereas the separate LSTM and neutrosophic methods achieved 87 and 78%, respectively. The model was incorporated into a mobile-centric architecture utilizing Fast Application Programming Interface, TensorFlow, Numerical Python, Structured Query Language + Alchemy to produce customized dietary recommendations. The results of this study contribute a proof-of-concept intelligent recommender for chronic kidney disease self-care, integrating sequential modeling and uncertainty management within a unified framework. The results are regarded as preliminary due to the evaluation being conducted in a simulated environment.

8 June 2026 in Engineering Proceedings
Verified
Conference Paper
Open access
PDF file
Dayron Rumbaut-RangelDayron Rumbaut-Rangel
Dayron Rumbaut-Rangel
FBFranklin Parrales Bravo
Franklin Parrales Bravo
RTRoberto Tolozano-Benites
Roberto Tolozano-Benites
LCLorenzo Jovanny Cevallos-Torres
Lorenzo Jovanny Cevallos-Torres
AI, specifically Bayesian networks, was applied in this study to diagnose and predict interruptions (whether due to faults or maintenance) that most significantly impact the time of interruption per kilowatt and frequency of maintenance intervention per kilowatt indicators in electrical distribution systems. Bayesian networks were employed to identify which functional stages, such as sub-transmission lines, distribution substations, and medium-voltage networks, exert the greatest influence on these performance metrics. Additionally, the analysis was conducted to categorize the interruption catalog and contribute to these impacts. By disaggregating the service quality and maintenance indicators reported monthly by the Guayas Los Ríos Business Unit of the National Electricity Corporation to Ecuador’s electricity sector regulatory bodies, the developed framework in this study enhances service reliability, optimizes maintenance planning, and reduces interruption times. Bayesian network models generated using R illustrate relationships between interruption causes and their impact on service quality indicators. Furthermore, a comparison of several models, including Naive Bayes, Tree Augmented Naive Bayes (TAN), and Backward Sequential Elimination and Joining (BSEJ), demonstrated that TAN and BSEJ achieved the highest accuracy in predicting interruption outcomes. These insights allow for more efficient targeting of maintenance resources, ultimately reducing the most impactful categories of interruptions and improving overall technical service quality.

21 October 2026 in Engineering Proceedings
Verified
Conference Paper
Open access
PDF file
LCLorenzo Jovanny Cevallos-Torres
Lorenzo Jovanny Cevallos-Torres
RCRosangela Caicedo-Quiroz
Rosangela Caicedo-Quiroz
Dayron Rumbaut-RangelDayron Rumbaut-Rangel
Dayron Rumbaut-Rangel
FBFranklin Parrales Bravo
Franklin Parrales Bravo
The COVID-19 pandemic exposed deficiencies in public health systems, particularly in prehospital emergency care. The increase in emergency contacts resulted in operational strain, markedly extended response times, and negative clinical consequences. To address the issues, we developed a machine-learning predictive model designed to estimate and optimize response time in prehospital clinical situations and aid dispatcher decision-making. Important variables were discerned by a scientific and documentary analysis, encompassing patient location, resource availability, event features, and hospital destination. A prototype was created utilizing a knowledge discovery in databases method that encompassed data preparation, model training, accuracy assessment, and model comparison. Supervised random forest and k-nearest neighbors (KNN) classifiers were employed to predict the optimal destination hospital and to automate aspects of the allocation process, hence minimizing delays. Results indicate that random forest surpasses KNN, attaining superior accuracy and facilitating quicker, more dependable decisions during peak demand periods. The model functions as an intelligent helper that enhances prehospital service efficiency and diminishes response times.