January 17 ~ 18, 2026, Zurich, Switzerland
Christian Hitz1, Vaclav Pechtor2, Ali Cem Guler1, Joel Egli1, Artan Perkola, 1Zurich University of Applied Sciences (ZHAW), Switzerland, 2Prague University of Economics and Business, Czech Republic
SMEs differ fundamentally from large organisations when it comes to succession planning. If large companies are looking for suitable promotions, SMEs are looking for external buyers who are difficult to find. Cultural differences between the previous owner and the successor, as well as resistance from the workforce, can also jeopardize succession. Inheritance disputes in family businesses are also legendary. This study carried out a survey on a family business in the wine industry that can be described as a small business with a production of 100,000 bottles per year. We know from this company that succession planning is considered to have failed and there is no buyer for the winery. One possible discussed solution to this problem could be a so-called decentralized autonomous organization, which has been researched in the sector with the WineDAO project, an innovative concept that explores the potential of a decentralized wine economy. The topic has been investigated by conducting a survey among wine experts, but also possible investors of a winery. A survey with 91 participants revealed a solid initial interest, with experienced investors over 45 years of age showing a greater willingness to invest. The findings highlight the potential of WineDAO but also underscore the need for strategic planning and a well-defined business model.
Succession Planning, DAO, Blockchain, Corporate Governance.
Ying-Jie Jheng, Department of Education, National Taiwan Normal University, Taiwan
The gaming industry constitutes an important part of the global economic system but also brings problems to a society and people. While the past studies analyzed the positive and negative impacts of gambling, research focusing on the feelings and educational expectations of dealers, especially from the educational angle, is scarce. The study, using Macao as an example, is devoted to exploring Macao dealers’ (as parents’) feelings and educational expectations on their children. According to the research findings, Macao dealers feel that they were seduced by the “money” and “surround people” to be dealers, and then become “alienated human beings.” Being paralyzed between “money” and “pride”, as the psychological state of “ambivalence”, the most important educational expectation for their children is “never repeat their alienated life.” Based on the results, some educational implications are addressed. First, in order to support dealers to fulfill their role of parents, the government should provide them with more educational resources from other institutions. Regarding school teachers, it is necessary to elevate educational expectations for students and dispel the vicious cycle that “not performing well at school leads to becoming a casino dealer.”
Macao, Dealer, Ambivalence, Educational Expectation.
Ibironke Oluwakemi1, Lawal Bukola2, Olayinka-Phillips Cynthia3, Ojo Olumuyiwa4, Olusola Philip4, Adekanmbi Oluyemi4 , Idowu Olufunmi4 1Faculty of Education and Communication, University of Lincoln, United Kingdom 2Department of Psychology, University of Ibadan, Nigeria 3School of Management, Cardiff Metropolitan University, United Kingdom 4Department of Office and Information Management, Lead City University Ibadan, Nigeria
Equality, diversity, and inclusion (EDI) remain vital components to achieving sustainable innovative practices in the Nigerian higher education system. This study explored the lived experiences of the teaching staff in the Office Technology and Management (OTM) department who are currently undergoing doctoral degrees in a private university in Ibadan, Nigeria. A qualitative research approach through in-depth interviews and thematic analysis was adopted for the study. A purposive sampling was also employed to recruit OTM teaching staff in the polytechnics undergoing their PhD programme in a private university. Four broad themes emerged from the findings of the study: previous experience syndrome, professional striving, and organizational and personal barriers. The findings revealed that, by prioritizing inclusivity, institutions can empower PhD students to generate sustainable and innovative solutions for managing information and office environments. Recommendations for academic institutions include funding research competence and offering targeted mentorship programmes, which are very vital for driving sustainability.
Doctoral Journey, Office technology & management staff, Sustainability, Equality, Diversity & Inclusion.
Harshita Pendli1 and Nenavath Srinivas Naik2 1B.Tech, Department of Computer Science and Engineering (AI & DS), Indian Institute of Information Technology Design and Manufacturing (IIITDM), Kurnool, India 2Associate Professor, Department of Computer Science and Engineering, Indian Institute of Information Technology Design and Manufacturing (IIITDM), Kurnool, India
Accident risk prediction plays a crucial role in improving the safety and reliability of intelligent transportation systems. This paper presents an end-to-end multimodal framework that analyses the driver’s behaviour, surrounding weather, and road surface to estimate the overall level of driving risk. The system integrates three deep-learning modules: a MobileNetV3-Large model for detecting driver drowsiness, an EfficientNet-B0 model for recognising weather scenes, and a YOLOv8n detector for identifying road hazards such as cracks, potholes, and open manholes. Each model produces a confidence score that is combined through a weighted fusion strategy to obtain a single risk value. Based on this value, the situation is classified into three categories—low, medium, or high risk. To make the result more interpretable, the fused outputs are processed by a locally running large language model (Ollama Mistral 7B), which generates short, descriptive text alerts. These alerts are synchronised with pre-recorded driver, weather, and road clips using the MoviePy framework to produce an offline dashboard video that visually shows the risk level along with the explanation. The proposed design demonstrates how deep learning, fusion, and language reasoning can together deliver an explainable and practical accident-warning system.
Accident Risk Prediction, Deep Learning, Multimodal Fusion, YOLOv8n, MobileNetV3-Large, EfficientNet-B0, Large Language Model, Ollama Mistral 7B, MoviePy, Dashboard Visualisation.
Jorge Amílcar Vizcaíno, People and Change Consultant, BSc Psychology, Buenos Aires, Argentina
The crisis of meaning in the modern era is characterized by constant change, existential anxiety, identity confusion, social isolation, and the fragmentation of traditional cultural myths and life narratives. These issues complicate the already challenging process of self-realisation, or individuation, as defined by Carl Jung. This study proposes Generative AI as a catalyst to support individuals throughout their life trajectories by integrating and modelling two distinct theoretical approaches: the Elliott Wave Principle (collective market analysis) and Jungian Archetypal Theory (depth psychology). This transdisciplinary framework is termed Individuation Wave Analysis (IWA). A Proof of Concept is presented involving biographical case studies of Carl Jung, Warren Buffett, and other historical and contemporary figures, analysed using DeepSeek AI. The study addresses the benefits of the IWA framework in enabling individuals to make more conscious and meaningful forward-looking decisions in their lifelong individuation journey, critically examines its limitations and potential for subjective bias, outlines necessary next steps for research and scientific validation, and identifies potential fields of application.
Individuation, Elliott Wave Principle, Jungian Archetypal Theory, Psychological Education, Generative AI, DeepSeek AI.
Muhammad Kashif Majeed, International Islamic university Malaysia, Malaysia
Generative Artificial Intelligence (AI) technologies such as ChatGPT are rapidly transforming global educational practices in the era of digital innovation. However, their pedagogical application within rural and religious educational settings remains underexplored. Grounded in the Technological Pedagogical Content Knowledge (TPACK) and Diffusion of Innovation (DOI) frameworks, this study investigates how generative AI influences pedagogical strategies among teachers in Islamic secondary schools in Southern Pakistan, particularly in the Kot Addu district. The research aims to: (1) determine the level of ChatGPT and AI adoption among teachers; (2) examine how generative AI impacts pedagogy, lesson planning, and student interaction; and (3) analyze gender-based differences in perceptions and utilization of these tools. A quantitative research design was employed using a structured questionnaire administered to 440 Islamic secondary school teachers. Data were analyzed using SPSS 2021 through descriptive statistics, t-tests, and correlation analysis. The findings reveal moderate adoption of AI technologies, which positively contribute to instructional innovation and teacher–student engagement. Notably, male teachers demonstrated slightly higher confidence and more active use of generative AI tools compared to female teachers. The study acknowledges limitations such as reliance on self-reported data, infrastructural constraints related to internet and device access, and cultural reservations regarding AI integration in religious education. The findings highlight the importance of targeted teacher training, culturally sensitive digital policies, and institutional support to ensure equitable integration of AI. Future research should incorporate qualitative approaches to capture deeper insights into teachers’ experiences, examine long-term impacts on learning outcomes, and explore policy-level enablers and barriers to AI integration in Islamic educational contexts.
ChatGPT, Generative AI, Pedagogy, Islamic schools, gender difference, south pakistan.
Muhammad Kashif Majeed, International Islamic university Malaysia, Malaysia
Generative Artificial Intelligence (AI) technologies such as ChatGPT are rapidly transforming global educational practices in the era of digital innovation. However, their pedagogical application within rural and religious educational settings remains underexplored. Grounded in the Technological Pedagogical Content Knowledge (TPACK) and Diffusion of Innovation (DOI) frameworks, this study investigates how generative AI influences pedagogical strategies among teachers in Islamic secondary schools in Southern Pakistan, particularly in the Kot Addu district. The research aims to: (1) determine the level of ChatGPT and AI adoption among teachers; (2) examine how generative AI impacts pedagogy, lesson planning, and student interaction; and (3) analyze gender-based differences in perceptions and utilization of these tools. A quantitative research design was employed using a structured questionnaire administered to 440 Islamic secondary school teachers. Data were analyzed using SPSS 2021 through descriptive statistics, t-tests, and correlation analysis. The findings reveal moderate adoption of AI technologies, which positively contribute to instructional innovation and teacher–student engagement. Notably, male teachers demonstrated slightly higher confidence and more active use of generative AI tools compared to female teachers. The study acknowledges limitations such as reliance on self-reported data, infrastructural constraints related to internet and device access, and cultural reservations regarding AI integration in religious education. The findings highlight the importance of targeted teacher training, culturally sensitive digital policies, and institutional support to ensure equitable integration of AI. Future research should incorporate qualitative approaches to capture deeper insights into teachers’ experiences, examine long-term impacts on learning outcomes, and explore policy-level enablers and barriers to AI integration in Islamic educational contexts.
ChatGPT, Generative AI, Pedagogy, Islamic schools, gender difference, south pakistan.
Lily Chen1 , Moddwyn Andaya2 1 Westridge School, 324 Madeline Dr, Pasadena, CA 91105 2 California State Polytechnic University, Pomona, CA 91768
MyMelody, a singing app created through Unity and C#, powered by an added element of AI, aims to become an alternative to traditional singing lessons, combating the limited accessibility, costliness, and other challenges people often face along their singing journey [2]. The three most important systems within the app are the Voice Type Identifier (for the user to identify their vocal range), Note Spawning (to display notes in the karaoke), and Song Scoring (for the user to determine their progress). The design process yielded many challenges, such as implementing math and outside algorithms to convert between pitch and frequency, making the UI fun and interactive for the user, and deciding how to score a user [1]. These challenges were addressed through iterative debugging and external research. System functionality was evaluated through a controlled test consisting of 10 predefined scenarios, of which the AI agent successfully handled 9. MyMelody is an innovative project that should be used increasingly, as it incorporates the power of artificial intelligence to help make music more accessible.
AI-Assisted Music Learning, Vocal Range Identification, Pitch Detection, Interactive Music Education
Da Xing1, Xinke Wang2, Marisabel Chang3, Yu Sun4, 1University College London, London, United Kingdom 2Chadwick School, Palos Verdes Peninsula, United States 3California State Polytechnic University, Pomona, Pomona, United States 4California State Polytechnic University, Pomona, Pomona, United States
Large language models (LLMs) have shown remarkable capabilities across diverse domains, yet their effectiveness as educational tutors remains underexplored, particularly regarding their adherence to pedagogically sound teaching methods. The Socratic method, emphasizing guided discovery through strategic questioning rather than direct instruction, represents a gold standard in educational practice. This paper introduces a comprehensive benchmark for evaluating LLMs’ ability to employ Socratic teaching methods in mathematics education. We present a systematic evaluation framework comprising 1,000 carefully curated mathematics questions spanning six topics and three difficulty levels, an automated student simulator for realistic conversation generation, and a multidimensional scoring system assessing direct answer avoidance, teaching quality, and correctness of guidance. Our empirical evaluation of three prominent LLMs (GPT-3.5, Claude 4 Sonnet, and Gemini 2.5 Flash) across this large-scale dataset reveals that modern LLMs demonstrate strong Socratic teaching capabilities, achieving overall scores above 9.7/10 across all evaluation dimensions. However, subtle differences emerge in teaching quality and performance across difficulty levels and topics. Our benchmark provides a replicable framework for assessing conversational AI tutoring systems and identifies key areas for improvement in automated Socratic pedagogy.
Socratic method, intelligent tutoring systems, large language models, educational technology, benchmark, mathematics education, conversational AI, pedagogical assessment
Satish Wagle1, Saroj Baral2, Jorge Vargas3, and Khem Poudel2 1Computational and Data Science, Middle Tennessee State University, Murfreesboro, TN, USA 2Department of Computer Science, Middle Tennessee State University, Murfreesboro, TN, USA 3Department of Engineering Technology, Middle Tennessee State University, Murfreesboro, TN, USA
Timely detection of anomalies is important in clinical data for identifying rare conditions, diagnostic errors, and healthcare efficiencies. This study proposed an unsupervised hybrid anomaly detection approach to combine clinical reports and chest X-rays using BioClinicalBERT and ResNet-18 embeddings. It integrates outlier filtering with Isolation Forest, applies a pruning step to retain the most anomalous candidates, and then performs local refinement using Local Outlier Factor. The pipeline was evaluated on 3,851 samples from the Open-I dataset. Model performance was explored with various outlier detection seeds, different pruned percentiles, and different LOF settings. The results indicated consistent and strong anomaly detection of outliers indices (121, 1632, 2265, 2673) that were identified as anomalies for most of the different configuration setting. The Jaccard similarity heatmap showed moderate to high overlap between results (scores between 0.4 and 0.8), and the t-SNE plots visually confirmed that different types of samples were well separated in space. IF-LOF modeling approach demonstrated evidence in capturing the multimodal physical patterns inherent in clinical data; images and medical text representations yet required no labelled data, and leveraged the benefits of global partitioning and local density-based refinement, complemented by globally semantic embeddings from BioClinicalBERT for reliable and valid anomaly detection in clinical practice.
Multimodal Anomaly Detection, Open-I Dataset, BioClinicalBERT, ResNet18, Unsupervised Learning, Isolation Forest, Local Outlier Factor, Pruning, t-SNE, Jaccard Similarity
Adiele Joshua Eze, Department of Computer Science and Informatics, Federal University Otuoke, Bayelsa State, Nigeria
Modern data mining systems face increasing demands for performance, scalability, and privacy preservation. As data volumes grow exponentially, platforms must evolve to support distributed architectures, real-time analytics, and secure processing. This paper presents a comprehensive study of current data mining platforms, evaluating their efficiency, scalability strategies, and privacy-preserving mechanisms. We propose a modular framework that integrates parallel processing, federated learning, and differential privacy to enhance system robustness. Experimental results on benchmark datasets demonstrate significant improvements in throughput and privacy compliance, offering a roadmap for next-generation data mining platforms.
Data Mining Platforms, Scalability, Efficiency, Privacy Preservation, Distributed Systems, Federated Learning, Differential Privacy, Big Data Analytics, Parallel Processing, and Secure Data Mining.
Audrey Han1 , Howard Lee2, 1 Walnut High School, 400 Pierre Rd, Walnut, CA 91789, 2California State University, Long Beach, 1250 Bellflower Blvd, Long Beach, CA 90840
Modern data mining systems face increasing demands for performance, scalability, and privacy preservation. As data volumes grow exponentially, platforms must evolve to support distributed architectures, real-time analytics, and secure processing. This paper presents a comprehensive study of current data mining platforms, evaluating their efficiency, scalability strategies, and privacy-preserving mechanisms. We propose a modular framework that integrates parallel processing, federated learning, and differential privacy to enhance system robustness. Experimental results on benchmark datasets demonstrate significant improvements in throughput and privacy compliance, offering a roadmap for next-generation data mining platforms.
Parkinso , Machine Learning, Python, Scikit-learn
John M. Acken1 , Naresh K. Sehgal2, 1 Portland State University, Portland, Oregon, 2Securing the Cloud, LLC, Santa Clara, California
One of the biggest challenges for an IT administrator in a private or public data centre is to ensure a fair usage of resources between different Virtual Machines (VMs). If a particular VM does excessive I/O or memory access, then other VMs running on that same server will experience a slowdown in their access of the same resource. This phenomenon is called noisy neighbours and results in a performance variation experienced by users over time, due to over consumption of shared resources by other VMs. In this paper, we look at real-life data showing the performance variability and present methods to detect it. A variation of this phenomenon is when a noisy neighbours can decipher confidential contents of a victim VM. We term it as a nosy neighbour, which can pose security risks. We conclude the paper with some ideas to prevent noisy or nosy neighbours.
Cloud, Performance, Multi-tenants, Virtual Machines, Security, Noisy Neighbour, Nosy Neighbour
Cláudio Ratke, Fabiano Oss, Giulia Larissa Conradi, Raquel Cristina Isensee da Silva, Robson Flávio Freitas and Kelvin Melim, Faculdade Senac Blumenau, Blumenau, Santa Catarina, Brasil
When developing software, it is essential to understand user needs and their interactions with the system. In this context, the user stories become tools used to transcribe functionalities and requirements of the software; in theory, this documentation should show the clients perspective in a clear and precise way. However, when written badly, they can cause misinterpretation and development issues that impact software quality, cause rework, and, in consequence, financial loss. Regarding this scenario, the proposal is to develop a system to analyse user stories using GPT-3 artificial intelligence, the same as that used by ChatGPT. The proposal presented in this project consists of the creation of a tool to diagnose and suggest corrections to grammatical, orthographic, and pattern errors in user stories written by the users in a web application that uses natural language in Portuguese and machine learning technology. This system is trained to recognize writing patterns and detect possible errors, applying the Requirements Smells concept and suggesting appropriate corrections that will improve the written quality and user experience when using the software. The collaboration with GPT-3 in this process is essential, given that it provides a set of advanced tools that guarantee its capability to work with a large variety of situations and writing contexts. In summary, the system of user story correction using GPT-3 artificial intelligence is an innovative solution capable of enhancing writing quality and improving the user experience.
User Stories, GPT-3, Agile Methodology.
David Kuhlen1, 1Technische Hochschule Lübeck, Mönkhofer Weg 239, 23562 Lubeck, Germany
Process automation plays a critical role in ensuring fast and efficient process execution at high order volumes. Consequently, the introduction of dropshipping also places extensive demands on the automation of business processes. While full process automation is technically feasible, it is subject to practical limitations. To this end, the dropshipping business model is analyzed in this paper with a focus on the technical implementation of automated business processes. The areas of action necessary for the implementation of automated dropshipping processes are analyzed. The analysis is based on a survey that examines the overall importance of the business model, the contribution of key action areas to developing integration solutions supporting automated processes, and future IT requirements. The study concludes that the DropShipping business model will gain increasing importance in the future. Furthermore, it became evident that requirements analysis, in particular, represents a significant effort while offering high value in developing integration solutions for automated process handling. The data analysis shows that the cost-benefit potential of using artificial intelligence methods to enable automated processes is optimal. Finally, it was found that creating automated DropShipping processes requires, to a particularly high degree, the expertise of software developers.
DropShipping, Process Automation, Software Engineering