IFIP Open Access Digital Library
Not a member yet
22614 research outputs found
Sort by
ALERT: A Framework for Efficient Extraction of Attack Techniques from Cyber Threat Intelligence Reports Using Active Learning
No full textPart 5: ML Attack, VulnerabilityInternational audienceIn the dynamic landscape of cybersecurity, curated knowledge plays a pivotal role in empowering security analysts to respond effectively to cyber threats. Cyber Threat Intelligence (CTI) reports offer valuable insights into adversary behavior, but their length, complexity, and inconsistent structure pose challenges for extracting actionable information. To address this, our research focuses on automating the extraction of attack techniques from CTI reports and mapping them to the MITRE ATT &CK framework. For this task, fine-tuning Large Language Models (LLMs) for downstream sequence classification shows promise due to their ability to comprehend complex natural language. However, fine-tuning LLMs requires vast amounts of annotated domain-specific data, which is costly and time-intensive, relying on the expertise of security professionals. To meet these challenges, we propose ALERT, a novel cybersecurity framework which leverages active learning strategies in conjunction with an LLM. This approach dynamically selects the most informative instances for annotation, thereby achieving comparable performance with a significantly smaller dataset. By prioritizing the annotation of samples that contribute the most to the model’s learning, our methodology optimizes the allocation of resources. As a result, our framework achieves comparable performance with a dataset that is 77% smaller, making it more efficient for extracting and mapping attack techniques from CTI reports to the ATT &CK framework
SmartSSD-Accelerated Cryptographic Shuffling for Enhancing Database Security
No full textPart 2: Crypto ApplicationInternational audienceGiven that databases often house sensitive and valuable data, ensuring data confidentiality and integrity of the databases is imperative. Encryption emerges as one of the predominant techniques employed in bolstering database security. Alongside encryption, shuffling also offers a viable approach to fortify the security of database. However, both encryption and shuffling requires huge amounts of system I/O requests which bring performance burden to the database server. In this paper, we propose a design to further enhance the security of shuffling algorithm and improve its efficiency by employing SmartSSD computational storage device from Samsung and AMD. We conduct experiments to evaluate the overhead of the improved effectiveness
For the Fun of IT– In Search for Sensemaking of Digitalization Training Program for Leaders in Schools and Pre-schools in Sweden
No full textInternational audienceThe local practice of digitalization in schools is a key component of digital inclusion and literacy policies in most states. However, despite all good ambitions of educational digitalization there are several challenges when it comes to implementation. This paper will elaborate on and contribute to the understanding of implementation of digitalization in governmental settings using an inductive approach with a specific focus on responsibility and sensemaking in mandatory versus voluntary implementation requirements.This paper presents a case study with a mixed-method approach to examine how a Swedish local government, in charge of public education, initiated and implemented a program to enhance the competence of all school leaders in leading digitalization. This program was undertaken despite differences in digital governance requirements in schools and preschools at the time. The case study illustrates the importance of connecting sensemaking theory in implementation processes to enhance the awareness of meaningful realization of digital government in local practices. Additionally, this paper highlights the methodological implications of such theoretical approaches
Navigating Privacy Regulations: Administrative Burden of Digital Self-Services for Vulnerable Citizens
No full textInternational audienceBalancing privacy regulations and usability is a challenge in the delivery of government welfare services through digital self-services. This study explores the experiences of vulnerable citizens with digital self-services in government welfare services, focusing on the administrative burdens of privacy regulation vulnerable citizens experience when trying to access and use the benefits for which they are eligible. Through interviews, focus groups and observations, three key costs - compliance, psychological, and learning – are examined. The findings shed light on how privacy regulation worsens the burden and usability of self-services and reproduce socioeconomic inequalities for vulnerable citizens by (1) limiting access to support, (2) putting them at risk of identity theft and fraud, and (3) demanding excessive documentation with limited opportunities to resubmit
Mitigating Administrative Burdens: Understanding the Role of Intermediaries in Co-producing Digital Self-services
No full textInternational audiencePublic sector organizations are increasingly adopting digital self-service solutions. With digital self-services, citizens are given a more active role in co-producing their services, as they serve themselves and perform tasks that were previously performed by public professionals. However, many citizens struggle with their expanding role as co-producers and experience burdens in their interaction with digital self-services. To mitigate these burdens, citizens often turn to so-called intermediaries for help. Intermediaries are third parties such as family members, friends as well as professionals that assist citizens during digital self-services or completely take over the responsibility from them. Although they are important co-producers, intermediaries have seldom been the focus of attention as their role during co-production is often invisible from the outside. We present two qualitative empirical studies from Norwegian and Brazilian welfare services. Our findings show the burdens citizens experience with digital self-services and illuminate how important personal intermediaries are to reduce citizens’ experience of burdens, resulting in triangulated or hybrid “co-production”
Towards a Conceptual Model for Enhancing Efficiency and Collaboration in Agile Ramp-Up Projects Planning
No full textPart 7: Collaborative Networks as Driver of Innovation in Organizations 5.0: ModelsInternational audienceAgility is a major concern for industries across manufacturing and service sectors for a variety of reasons. For instance, frequent product development to meet evolving customer requirements and changing markets requires efficient and agile processes. Therefore, companies need to capitalize on product and service development and ramp-up projects in order to improve their efficiency as well as time-to-market. This paper provides preliminary results to address this gap by proposing a model covering conceptual underpinnings of agile ramp-up management. The model is developed iteratively following the design science research methodology. A simple excel prototype resulting from the model is developed in order to support agile ramp-up project planning. The model as well as the prototype are expected to help decision makers in scoping ramp-up projects and in improving the consistency of these projects planning. They also support the collaboration within ramp-up project team through sharing a common understanding of the project scope
Associations Between Gender Attributions and Social Perception of Humanoid Robots
No full textPart 2: Human-Machine CollaborationInternational audienceWith the thriving integration of robots in work spaces, user acceptance and trust in robots are particularly important. Both aspects are influenced by various factors, e.g., social perception. Gender considerably affects social perception of humans, however, whether gender attributions to robots hold similar implications is still unclear. We investigated this question with two samples (N1 = 238, N2 = 133) who rated four images of humanoid robots in terms of gender and social dimensions, i.e., anthropomorphism, sociability/morality, activity/cooperation, and competence. We found that in both samples gender perception differed significantly, but only perceived competence and sociability/morality varied as a function of gender: More femininity was associated with higher attributions of sociability/morality and lower attributions of competence. We take this as an indicator that gender influences social perception and should be considered as additional aspect when it comes to designing robots. However, gendering robots might also pose ethical risks in terms of deception and unwanted amplification of gender stereotypes
Continual Learning for Human-Machine Collaboration in VUCA Environments
No full textPart 2: Human-Machine CollaborationInternational audienceThis study presents a novel approach to enhancing human-machine collaboration (HMC) in volatile, uncertain, complex, and ambiguous (VUCA) environments by emphasizing the importance of continual learning. Addressing the limitations of traditional static systems, the proposed HMC system integrates continual learning and object detection algorithms to enhance error handling, operational efficiency, and resilience. The research aims to establish a new standard for intelligent HMC systems, emphasizing ongoing reciprocal learning between humans and machines to improve decision-making and performance. Practical implementation demonstrates the system’s effectiveness in reducing downtime and increasing adaptability. By integrating human expertise and machine intelligence, the system fosters improved problem-solving capabilities and operational efficiency, making it highly suitable for dynamic and unpredictable industrial settings. This study addresses critical gaps in current methodologies, providing a comprehensive framework for the future of HMC in complex manufacturing environments
Do There Exist an Emotion Trend in Scientific Papers? PRO-VE Conference as a Case
No full textPart 3: Emotions and Collaborative NetworksInternational audienceScientific writing aims for formality and objectivity, yet emotions are integral to human communication, decision-making, and collaboration, all of which are fundamental to scientific progress. Existing research on emotion detection has mainly focused on datasets from social media and online platforms, where emotional expressions are abundant. However, scientific texts pose unique challenges due to their formal language and the rarity of explicit emotional words, necessitating specialised investigation. This study investigates the presence and nature of emotions in scientific texts, specifically analysing the abstracts from the PRO-VE conference series from 2012 to 2022. Two emotion detection methods are employed: a lexicon-based approach and a hybrid machine learning-based approach. The lexicon-based approach utilises the NRC Emotion Lexicon to identify and quantify emotions within the PRO-VE abstracts, while the hybrid approach integrates Word2Vec for word embedding generation and a Random Forest classifier for emotion prediction. The findings reveal a predominance of positive emotions, such as trust, anticipation, and joy, in the PRO-VE abstracts, consistent with the objective nature of scientific writing. In light of the PRO-VE conference series’ 25th anniversary, an analysis of trends and patterns in the detected emotions offers insights into the emotional landscape of this prestigious conference series. The study also critically examines the limitations of the experiments, including the dataset size and the prevalence of positive emotions
Emotions in Human-AI Collaboration
No full textPart 3: Emotions and Collaborative NetworksInternational audienceThis exploratory paper addresses the role of emotions in the management of collaborative networks (CNs) amid the rise of hybrid teams consisting of humans and AI agents. Building on previous research that emphasizes the critical role of emotions in fostering trust and preventing conflicts within CNs, we propose expanding these emotional frameworks to accommodate hybrid collaborative networks. The paper reviews the significance of human-AI collaboration, highlighting the complementary strengths of both and identifying three research streams: affective/sentient AI agents, human emotions modelling, and collective hybrid network emotions. Emphasizing the underexplored area of collective emotions, we suggest leveraging these insights to enhance the management and sustainability of hybrid networks. A framework for emotions estimation in CNs is described. Our aim is to identify challenges and guide future research in the integration of emotional intelligence within human-AI collaborative environments