Conference Agenda

Abstract Twin green-digital transitions require engineers who can reason about environmental impacts and make defensible sustainability decisions. This paper reviews recent work on artificial intelligence (AI) and immersive virtual reality (IVR) in higher education, with a focus on ecology-related learning. A qualitative literature review was conducted across major academic databases to identify current approaches, reported benefits, and governance risks. Building on these findings, we present EcoAI, a course-scoped AI ecology consultant delivered through Moodle (LTI) as a Streamlit application. EcoAI structures student work in four steps—problem framing, causal reasoning, trade-off analysis, and reporting with verification—and produces assessable artefacts such as checklists, causal chains, and trade-off matrices. To support responsible use, the system enforces guardrails that require explicit assumptions, mechanism-based justification, trade-offs, and verification. The combined review-and-implementation contribution outlines a practical pathway for using AI to strengthen sustainable thinking in engineering ecology education while addressing reliability and accountability concerns.

Universities across developing countries face persistent challenges in delivering reliable e-learning due to constrained network infrastructure. Software-Defined Networking (SDN) has demonstrated potential to improve traffic management, bandwidth allocation, and security in such environments; however, real-world adoption remains limited. This paper presents a systematic literature review following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) methodology, analysing 26 peer-reviewed studies published between 2018 and 2025. Nineteen recurring challenges are identified across technical, financial, human resource, and governance dimensions. A review-derived institutional readiness framework is synthesised from the included studies, and three practical deployment pathways are proposed, each aligned with a distinct institutional resource profile. The findings provide an evidence-based foundation for network administrators, institutional decision-makers, and policymakers seeking to enable scalable and sustainable SDN-supported e-learning infrastructure in resource-constrained environments.

This paper presents an integrated pedagogical model for the formation of competencies in students in the field of machine learning and its application in the valorization of residual products from food production. The proposed methodology combines real multisensory experimental data from biscuits enriched with aronia, blackcurrant and goji berry flours with a sequential training algorithm, including pre-processing, feature selection, PCA/LV analysis, building and optimization of ML models. The developed 12-step training process overcomes pedagogical and didactic challenges related to the perishability of food samples, data variability and limited laboratory time, by using standardized arrays and virtual ML laboratories. The results show that the integration of ML in training increases student engagement, develops analytical and critical thinking and supports the acquisition of skills for modeling, interpretation and optimization of food processes. The model demonstrates high applicability in the context of the circular economy and the digital transformation of food technologies, offering a sustainable framework for training future specialists in the subject area.

Urban disasters impose post-traumatic stress dis- order (PTSD)-like burdens on youth. However, IT-supported resilience interventions remain scarce. Limited scalable AI- empowered solutions combining emotional support, engagement theory, and service-learning for higher education post-trauma recovery. We recruit 48 university students (SUs) to join an AI-empowered train-the-trainer (AITT) program. This program includes multi-stage train-the-trainer activities, such as positive seminars and AI-empowered social robot workshops. The train- ers instruct primary students who are affected by the Tai Po fire. We employ pre- and/or post-questionnaires of self-determination theory (SDT), Statistical Manual of Mental Disorders, 5th Edition (DSM-5), and teaching and learning. Engagement and motivation demonstrate statistically significant improvement. Teaching efficacy and learning experience exhibit moderate to high effectiveness. The SUs report that they achieve interpersonal success. To conclude, the AITT program builds resilience and competencies. This is scalable for IT-based higher education disaster response.