IFIP Open Access Digital Library
Not a member yet
22614 research outputs found
Sort by
Self-harm Detection from Texts: A Comparative Study Utilizing BERT, Machine Learning, and Deep Learning Approaches
No full textPart 1: Applications of AI/ML in Natural Language ProcessingInternational audienceIn a rapidly evolving world, the demands of modern life contribute to rising global anxiety and depression. Mental health, often neglected due to stigma, leads to untreated issues, including self-harm and suicide. Social media has become a platform for expressing mental health concerns, and employing suitable algorithms enables automated suicide sentiment detection. This research compares various BERT models to identify an efficient approach for leveraging social media to facilitate professional help-seeking without stigma. The study compared outcomes from different BERT models against those from conventional methods including Logistic Regression and Random Forest. The study also extended to incorporate deep learning models, specifically CNN, LSTM, and BiLSTM, offering a comprehensive analysis across diverse approaches to assess their effectiveness in text classification tasks. Through these experiments, we attained exceptional F1 scores of 99% for RoBERTa, 98% for AlBERT, and 96% for BERT base. In contrast, traditional models like Logistic Regression achieved 93%, Random Forest 89%, and deep learning models such as LSTM, BiLSTM and CNN achieved 82%, 93% and 90%, respectively. The results of these models are compared with each other to study and draw inferences on their performances. These findings hold significant implications for the development of more robust and language-specific Suicide text detection systems, contributing to the overall effort to curb the increasing suicide rates all over the world
Anticipating Future College Admission Cutoffs: An Innovative Predictive Model Incorporating Student Reviews and Historical Admissions Cutoff Data Using Machine Learning
No full textPart 1: Applications of AI/ML in Natural Language ProcessingInternational audienceIn this era of ever-increasing competition for the higher education opportunities, the capability to predict future college admission cutoff ranks plays an important role in shaping the educational journeys of prospective students. This work presents an innovative predictive model aimed at assisting students in making well-informed decisions when it comes to college selection. The proposed framework harnesses the power of key elements: student-personal reviews and historical admission cutoff, National Institutional Ranking Framework (NIRF) factors. This predictive model combines the student reviews, with the insights of admission cutoff ranks and NIRF scores. This work showcases a comprehensive approach that not only anticipates future admission cutoffs but also empowers students to make enlightened enrolment choices. This study explores the intricate interplay between student feedback and historical trends, providing a comprehensive perspective on the evolving landscape of college admissions
Neuro-Evolution-Based Language Model for Text Generation
No full textPart 1: Applications of AI/ML in Natural Language ProcessingInternational audienceIn the dynamic field of natural language processing, the enhancement of text generation models presents a complex challenge, compounded by the intricate architectures and substantial parameters of contemporary neural networks. This study introduces a groundbreaking method that applies Genetic Algorithms to evolve the architecture of Long Short-Term Memory networks (LSTM), specifically tailored for text generation tasks. Our approach employs a sophisticated gene encoding mechanism that captures the diverse LSTM network configurations and the optimal depth of the network required for generating coherent and contextually relevant text. Our method systematically refines and evolves network architectures through iterative selection, crossover, and mutation processes, uncovering the most effective configurations for text generation. This evolutionary process is designed to yield an LSTM network architecture with enhanced performance in text generation
Towards the Integration of Conversational Agents Through a Social Media Platform to Enhance the Agility of BPM
No full textPart 1: AI and CollaborationInternational audienceBusiness Processes enable collaboration among various stakeholders, allowing different groups (people, organizations) to work together to achieve common goals. Therefore, optimizing Business Process Management (BPM) is essential for organizational success in today’s dynamic business environment. However, traditional BPM methods often struggle in volatile execution environments characterized by rapid change, dynamic customer demands, and evolving market trends. Innovative strategies are needed to enhance BPM practices and increase the agility of collaborative business processes. To this end, a particularly promising approach is to use Large Language Models (LLM) agents (Artificial Intelligence conversational agents). These AI conversational agents can be integrated into a social media platform to ease the stakeholders’ collaboration by supporting the co-construction, design, modification, execution, and monitoring of collaborative business processes. AI conversational agents in social media platforms democratize BPM by facilitating collaborative process design and execution, streamlining interactions, and fostering seamless communication and personalized assistance, thus enhancing agility
Managing Risks in Collaborative Network Organizations Within Sales and Operations Planning: A Maturity Model
No full textPart 7: Collaborative Decision Making 303International audienceSales and Operations Planning (S&OP) is essential for aligning strategic plans with daily operations. However, the dynamic nature of modern Collaborative Network organizations presents challenges due to uncertainties and risks. This research aims to fill this gap by creating a maturity model to manage uncertainties in the S&OP process. A literature review was conducted, examining six key dimensions: Process, Tools, People, Objectives, Decisions, and Key Performance Indicators (KPIs). The review uncovered gaps and limitations in current research on uncertainty management in S&OP, underscoring the need for a maturity model for Collaborative Network organizations. As a result, a three-stage maturity model was developed to evaluate Collaborative Network organizations’ S&OP practices, providing guidance to manage uncertainties within the S&OP process. The study highlights the need for future research to develop well-defined procedures for effectively addressing uncertainties, including scenario planning methodologies and decision-making frameworks that account for uncertainties
Leveraging Sentiment Analysis and Reporting for Re-designing Business Processes Using Large Language Models: A SentiProMo Case Study in Airline Check-In Processes
No full textPart 1: Digital Transformation Approaches in Production and ManagementInternational audienceIn today’s competitive landscape, optimizing business processes is crucial for maintaining efficiency and customer satisfaction, particularly in the manufacturing sector. This paper presents SentiProMo, a self-developed tool that integrates sentiment analysis of collaborative comments with summarization features to enhance the design phase of business processes. Leveraging the capabilities of large language models (LLMs) such as Chat-GPT, this tool empowers managers with insightful reports for informed decision-making. To demonstrate the effectiveness of SentiProMo, a case study was conducted focusing on the check-in process at airports, a critical aspect of the airline industry. Real-world data from collaborative comments during the design phase of the check-in process were analyzed using sentiment analysis techniques. Additionally, summarization features were employed to generate concise and informative reports for management stakeholders. The results of the case study showcase the potential of LLM-powered tools like SentiProMo in streamlining business processes. By harnessing sentiment analysis, organizations can gain valuable insights into employee perceptions and identify areas for improvement. Moreover, the summarization capabilities facilitate the efficient communication of findings to management, enabling them to make informed decisions promptly. This research not only underscores the power of LLMs in enhancing business process design but also highlights the promising avenues for future applications. With its wide array of potential applications across various industries, SentiProMo represents a significant advancement in process optimization and management
Ask DoctorBot: Unpacking the Social Design Aspects of Symptom Assessment Chatbots
No full textPart 1: Artificial Intelligence Adoption and ImpactInternational audienceConversational AI applications or chatbots have raised fundamental questions on understanding emerging user experiences. Healthcare chatbots offer omnipresent and responsive support for symptom checking by acting as virtual health assistants. The research focuses on the social design aspects of symptom assessment chatbots - in particular - emotional support and automated social presence. A survey study with 223 users of a symptom assessment chatbot in China supports the significant impact of these two design elements on perceptions of functional service quality. However, the same effect was not found for service satisfaction, suggesting a disconnection between interaction elements and expectations about what symptom assessment chatbots can achieve for users. Reflecting on the results, the research identifies the influence mechanisms of healthcare chatbots’ social design on the processes and outcomes of user engagement
Using Information Technology to Create Sustainable Organizational Constellations
No full textPart 2: Digital Transformation and Organizational InnovationInternational audienceOrganizations build alliances to create products or services that fulfill a specific market need. These alliances have been around for a while, and researchers describe them in various terms like Virtual Organization, or Ecosystem. The manifestation of these alliances increased due to the increasing interconnection between organizations within the fourth industrial revolution. The never-ending alignment between markets and alliances provides a dynamic landscape for organizations and their partners. While abundant research is available on these constellations, we do not know what IT needs can be used to cope with these dynamics. Our research aims to fill that gap, specifically on the relations between network participants and their interactions. We executed a multi-case study on 10 alliances comprising 35 interviews, which were transcribed and coded. By analyzing the co-occurrence, we could determine relationships between the organizational characteristics and the IT needs. Our results validated the characteristics framework, provided a list of IT needs, and showed that IT needs can positively influence and strengthen organizations to cope with market dynamics
A Bitcoin-Based Digital Identity Model for the Internet of Things
No full textInternational audienceCybersecurity in the Internet of Things (IoT), at its heart, relies on the digital identity concept to build security mechanisms such as authentication and authorization. However, current centralized identity management systems are built around third party identity providers, which raises privacy concerns and presents a single point of failure. In addition, IoT unconventional characteristics such as scalability, heterogeneity and mobility require new identity management systems to operate in distributed and trustless environments. In order to deal with these challenges, we present the Blockchain-based Identity Management System for the Internet of Things. By such, things and people are able to self-manage their identities and authenticate without relying on any third parties
Security Challenges and Countermeasures in Blockchain’s Peer-to-Peer Architecture
No full textInternational audienceThis paper addresses the issue of security in blockchain systems, with a focus on attacks targeting the peer-to-peer architecture. The peer-to-peer nature of blockchain is fundamental to many of the benefits promised by blockchain applications. We detail various attacks affecting this architecture, including network attacks, eclipse attack, majority attacks, selfish mining attack, block-withholding attacks, and time-jacking attack. This paper provides a significant contribution in three parts: firstly, it offers a comprehensive description of several attacks targeting this architecture. Secondly, it examines the necessary conditions for the effectiveness of these attacks and, thirdly, it presents a qualitative overview of the defense strategies identified in the existing literature to deal with these threats. © IFIP International Federation for Information Processing 2024