Most expert projections indicate that in 2030, there will be over one billion people aged 60 or over. The vast majority of them prefer to spend their last years at home, and almost a third of them live alone. This creates a growing need for technology-based solutions capable of helping older people to live independently in their places. Despite the wealth of solutions proposed for this general problem, there are very few support systems that can be reproduced on a larger scale. In this study, we propose a method to monitor the activity of the elderly living alone and detect deviations from the previous activity patterns based on the idea that the residential living environment can be modeled as a collection of behaviorally significant places located arbitrarily in a generic space. Then we use virtual pheromones—a concept defined in our previous work—to create images of the pheromone distribution maps, which describe the spatiotemporal evolution of the interactions between the user and the environment. We propose a method to detect deviations from the activity routines based on a simple statistical analysis of the resulting images. By applying this method on two public activity recognition datasets, we found that the system is capable of detecting both singular deviations and slow-deviating trends from the previous activity routine of the monitored persons.

(1) Background and objectives: Global crises such as climate change, pandemics, and resource depletion present unprecedented challenges that require coordinated and innovative responses. This study advocates for a bottom-up strategy by proposing and developing tools that empower individuals and organizations to embrace green and sustainable entrepreneurship, fostering localized solutions with global impact. In this context, the research explored the integration of technology-enabled entrepreneurship education (EE) into engineering curricula, with a focus on sustainable and green entrepreneurship. Leveraging insights from European projects, the research developed a cloud-based entrepreneurship educational system featuring three e-learning platforms. Particularly, one e-learning platform was specifically designed to support green and sustainable entrepreneurship education. (2) Methods: the study involved literature reviews, stakeholder surveys, interviews with green entrepreneurs, and iterative platform design. (3) Results: Key outcomes of the research include a methodology for designing pedagogical strategies and learning content for green and sustainable EE. Additionally, digital tools such as the green business innovation canvas, TRIZ-inspired matrices, and AI-driven recommendation systems were created to equip learners with the skills to develop sustainable business models. (4) Originality: The study’s novelty lies in its integration of advanced digital tools with original pedagogical strategies, providing a scalable framework for incorporating sustainability into entrepreneurship education. These findings have practical implications for educators and policymakers working to promote eco-friendly business practices. Future research should investigate the scalability of these tools across diverse educational contexts and evaluate their long-term impact on fostering sustainable entrepreneurial mindsets.

The evaluation of large language models (LLMs) increasingly relies on other LLMs acting as automated judges. While this approach offers scalability and efficiency, it raises serious concerns regarding evaluator reliability, positional bias, and ranking stability. This paper presents a scalable framework for diagnosing positional bias and instability in LLM-based evaluation by using controlled pairwise comparisons judged by multiple independent language models. The system supports mirrored comparisons with reversed response order, prompt injection, and surface-level perturbations (e.g., paraphrasing, lexical noise), enabling fine-grained analysis of evaluator consistency and verdict robustness. Over 3600 pairwise comparisons were conducted across five instruction-tuned open-weight models using ten open-ended prompts. The top-performing model (gemma:7b-instruct) achieved a 66.5% win rate. Evaluator agreement was uniformly high, with 100% consistency across judges, yet 48.4% of verdicts reversed under mirrored response order, indicating strong positional bias. Kendall’s Tau analysis further showed that local model rankings varied substantially across prompts, suggesting that semantic context influences evaluator judgment. All evaluation traces were stored in a graph database (Neo4j), enabling structured querying and longitudinal analysis. The proposed framework provides not only a diagnostic lens for benchmarking models but also a blueprint for fairer and more interpretable LLM-based evaluation. These findings underscore the need for structure-aware, perturbation-resilient evaluation pipelines when benchmarking LLMs. The proposed framework offers a reproducible path for diagnosing evaluator bias and ranking instability in open-ended language tasks. Future work will apply this methodology to educational assessment tasks, using rubric-based scoring and graph-based traceability to evaluate student responses in technical domains.

Background and objectives: Large language models (LLMs) show promise in automating open-ended evaluation tasks, yet their reliability in rubric-based assessment remains uncertain. Variability in scoring, feedback, and rubric adherence raises concerns about transparency and pedagogical validity in educational contexts. This study introduces CourseEvalAI, a framework designed to enhance consistency and fidelity in rubric-guided evaluation by fine-tuning a general-purpose LLM with authentic university-level instructional content. Methods: The framework employs supervised fine-tuning with Low-Rank Adaptation (LoRA) on rubric-annotated answers and explanations drawn from undergraduate computer science exams. Responses generated by both the base and fine-tuned models were independently evaluated by two human raters and two LLM judges, applying dual-layer rubrics for answers (technical or argumentative) and explanations. Inter-rater reliability was reported as intraclass correlation coefficient (ICC(2,1)), Krippendorff’s α, and quadratic-weighted Cohen’s κ (QWK), and statistical analyses included Welch’s t tests with Holm–Bonferroni correction, Hedges’ g with bootstrap confidence intervals, and Levene’s tests. All responses, scores, feedback, and metadata were stored in a Neo4j graph database for structured exploration. Results: The fine-tuned model consistently outperformed the base version across all rubric dimensions, achieving higher scores for both answers and explanations. After multiple-testing correction, only the Generative Pre-trained Transformer (GPT-4)—judged Technical Answer contrast remains statistically significant; other contrasts show positive trends without passing the adjusted threshold, and no additional significance is claimed for explanation-level results. Variance in scoring decreased, inter-model agreement increased, and evaluator feedback for fine-tuned outputs contained fewer vague or critical remarks, indicating stronger rubric alignment and greater pedagogical coherence. Inter-rater reliability analyses indicated moderate human–human agreement and weaker alignment of LLM judges to the human mean. Originality: CourseEvalAI integrates rubric-guided fine-tuning, dual-layer evaluation, and graph-based storage into a unified framework. This combination provides a replicable and interpretable methodology that enhances the consistency, transparency, and pedagogical value of LLM-based evaluators in higher education and beyond.

(1) Background and objectives: Large language models (LLMs) such as GPT, Mistral, and LLaMA exhibit strong capabilities in text generation, yet assessing the quality of their reasoning—particularly in open-ended and argumentative contexts—remains a persistent challenge. This study introduces Dialectical Agent, an internally developed modular framework designed to evaluate reasoning through a structured three-stage process: opinion, counterargument, and synthesis. The framework enables transparent and comparative analysis of how different LLMs handle dialectical reasoning. (2) Methods: Each stage is executed by a single model, and final syntheses are scored via two independent LLM evaluators (LLaMA 3.1 and GPT-4o) based on a rubric with four dimensions: clarity, coherence, originality, and dialecticality. In parallel, a rule-based semantic analyzer detects rhetorical anomalies and ethical values. All outputs and metadata are stored in a Neo4j graph database for structured exploration. (3) Results: The system was applied to four open-weight models (Gemma 7B, Mistral 7B, Dolphin-Mistral, Zephyr 7B) across ten open-ended prompts on ethical, political, and technological topics. The results show consistent stylistic and semantic variation across models, with moderate inter-rater agreement. Semantic diagnostics revealed differences in value expression and rhetorical flaws not captured by rubric scores. (4) Originality: The framework is, to our knowledge, the first to integrate multi-stage reasoning, rubric-based and semantic evaluation, and graph-based storage into a single system. It enables replicable, interpretable, and multidimensional assessment of generative reasoning—supporting researchers, developers, and educators working with LLMs in high-stakes contexts.

(1) Background and objective: Mobility is crucial for healthy aging, and its loss significantly impacts the quality of life, healthcare costs, and mortality among older adults. Clinical mobility assessment methods, though precise, are resource-intensive and economically impractical, and most of the existing solutions for automatic detection of mobility anomalies are either obtrusive or improper for long time monitoring. This study explores the feasibility of using non-intrusive, low-cost binary sensors for continuous, remote detection of mobility anomalies in older adults, aiming to identify both sudden mobility events and gradual mobility loss. (2) Method: The study utilized publicly available datasets (CASAS Aruba and HH120) containing annotated activity data recorded from binary sensors installed in residential environments. After data preprocessing—including filtering irrelevant sensor events and aggregation into behaviorally meaningful places (BMPs)—a time series forecasting model (Prophet) was used to predict normal mobility patterns. A fuzzy inference module analyzed deviations between observed and predicted sensor data to determine the probability of mobility anomalies. (3) Results: The system effectively identified periods of prolonged inactivity indicative of potential falls or other mobility disruptions. Preliminary evaluation indicated a detection rate of approximately 77–81% for point mobility anomalies, with a false positive rate ranging from 12 to 16%. Additionally, the approach successfully detected simulated gradual declines in mobility (1% per day reduction), evidenced by statistically significant regression trends in activity levels over time. (4) Conclusions: The study argues that non-intrusive binary sensors, combined with lightweight forecasting models and fuzzy inference, may provide a practical and scalable solution for detecting mobility anomalies in older adults. Although performance can be further enhanced through improved data preprocessing, predictive modeling, and anomaly threshold tuning, the proposed system effectively addresses key limitations of existing mobility assessment approaches.