Conference Agenda

The rapid adoption of large language models (LLMs), such as ChatGPT, has introduced significant tensions in the assessment of mathematics in higher education, especially in engineering and science programs. Although these models demonstrate a high capacity for solving mathematical problems and generating coherent explanations, their integration challenges the validity of traditional assessment practices focused on the final product. This study presents a critical review of the literature on the use of LLMs in university mathematics assessment, based on a corpus of 27 articles indexed in Scopus, published between 2023 and 2026. A critical analysis methodology was employed, based on categories of pedagogical alignment, validity of evidence, assessment transformation, academic risks, and future projections. The results show that most studies use LLMs primarily as solution generators, while assessment instruments remain largely unchanged. Exam- and task-based assessments predominate, with little attention paid to process assessment, argumentation, or transfer. Likewise, methodological limitations in the empirical evidence and a predominantly declarative treatment of risks such as academic integrity and cognitive dependency are identified. Overall, the review highlights the need for a redesign of mathematics assessment that prioritizes evidence of reasoning, explanation, and verification, integrating LLMs in a critical and pedagogically aligned manner to preserve rigor, equity, and validity in higher education.

Calculus teaching in engineering in Latin America faces challenges such as high failure rates and student disconnection. The study "More Than Points and Badges" analyses the impact of narrative gamification through a mixed-methods approach. Quantitatively, the "Path of the Integrator" intervention was evaluated in integration techniques using a quasi-experimental design. Qualitatively, student perceptions and emotional connections were explored.

The results show high acceptance and a positive perception of the ludic methodology. Findings highlight a 95.3% increase in intrinsic interest and an 82% reduction in academic stress, boosting motivation and engagement. However, critical challenges were identified: 48% of participants were initially unfamiliar with the methodology, emphasizing the need for an introductory phase to establish the "ludic contract". Additionally, the importance of guaranteeing equitable conditions for all students and adjusting the feedback mechanism was detected, as 22% did not consider it useful for their formative process.

In conclusion, narrative gamification is a powerful tool for transforming mathematics learning. Its optimal implementation requires clear socialization, an inclusive instructional design that mitigates negative competition, and a personalized feedback system.

This study develops the calibration and validation of the Concrete Damaged Plasticity (CDP) model for the nonlinear finite element simulation of reinforced concrete beams subjected to monotonic bending. The research integrates a constitutive formulation based on continuous damage plasticity, two-dimensional discretization under plane stress conditions, and an iterative parametric fitting procedure implemented in Abaqus/Standard. The methodology was structured in four stages: (i) geometric definition and reinforcement modeling using embedded region techniques, (ii) constitutive characterization of the concrete, including Hognestad-type compression curves and tensile softening regularized by fracture energy, (iii) calibration of critical parameters of the CDP model—dilatance angle, regularization viscosity, and damage parameters—and (iv) quantitative validation by comparison with experimental results reported in the literature.

The rapid incorporation of generative artificial intelligence into higher education has opened up new possibilities for teaching mathematics in engineering programs; however, its unstructured use poses risks associated with superficial learning, cognitive dependency, and lack of conceptual transfer. Despite the growing volume of studies on generative AI in education, there remains a gap in terms of explicit pedagogical frameworks that guide its integration toward deep mathematical learning. In this context, the present study aims to design and conceptually ground a pedagogical framework to promote deep mathematical learning through the use of generative artificial intelligence in engineering education. The research adopts a qualitative approach with a non-experimental theoretical-conceptual design, based on an integrative synthesis and critical analysis of recent scientific literature. As a result, a framework structured in five interrelated layers is proposed: deep mathematical learning intentions, pedagogical roles of generative AI, human-AI interaction patterns, didactic sequences oriented towards deep reasoning, and criteria for authentic assessment and ethical management. The framework emphasizes the role of AI as a regulated cognitive mediator, subordinate to pedagogical and metacognitive principles. It is concluded that the proposal constitutes a solid conceptual basis for guiding future research aimed at the implementation and empirical validation of the pedagogical use of generative AI in the teaching of mathematics in engineering.

This study aimed to determine the relationship between the STEAM methodology and the learning of differential equations in Calculus III students at FIIS-UNAC, 2025. An applied quantitative investigation with a non-experimental, correlational and cross-sectional design was conducted, using a census sample of 30 students. A Likert-type questionnaire on STEAM methodology and a differential equations learning test were administered, both validated by expert judgment and showing high reliability. Because the data did not follow a normal distribution (Shapiro–Wilk), Spearman’s Rho was used. The findings revealed very strong positive correlations between STEAM and learning (ρ = 0, 972; p < 0, 001), as well as with academic performance, conceptual understanding and practical application in this group of future engineers within an engineering faculty, thereby confirming the proposed hypotheses.

We propose the Santiago Dynamical System (SDS) as a unifying formulation of discrete dynamics on N generated
by a coding and an observable. The central case studied in this work is the length SDS, where the observable is the length of
a word α(n) over an alphabet Σ, giving rise to the iteration nk+1 = |α(nk)|. This perspective connects functional graphs on N
with symbolic dynamics, coding theory, and Lyapunov-type potential methods. Universal structural properties are presented,
including contraction/expansion criteria and a methodological framework to characterize attractors, basins, and iterative
complexity in classes of natural codings. Finally, conjectures are formulated concerning small attractors and iterated-logarithmtype
convergence in compact codings.