IFIP Open Access Digital Library
Not a member yet
22614 research outputs found
Sort by
Data and Applications Security and Privacy XXXVIII: 38th Annual IFIP WG 11.3 Conference, DBSec 2024, San Jose, CA, USA, July 15–17, 2024, Proceedings
No full textInternational audienceBook Front Matter of LNCS 1490
Does Differential Privacy Prevent Backdoor Attacks in Practice?
No full textPart 7: Differential PrivacyInternational audienceDifferential Privacy (DP) was originally developed to protect privacy. However, it has recently been utilized to secure machine learning (ML) models from poisoning attacks, with DP-SGD receiving substantial attention. Nevertheless, a thorough investigation is required to assess the effectiveness of different DP techniques in preventing backdoor attacks in practice. In this paper, we investigate the effectiveness of DP-SGD and, for the first time, examine PATE and Label-DP in the context of backdoor attacks. We also explore the role of different components of DP algorithms in defending against backdoor attacks and will show that PATE is effective against these attacks due to the bagging structure of the teacher models it employs. Our experiments reveal that hyper-parameters and the number of backdoors in the training dataset impact the success of DP algorithms. We also conclude that while Label-DP algorithms generally offer weaker privacy protection, accurate hyper-parameter tuning can make them more effective than DP methods in defending against backdoor attacks while maintaining model accuracy
Understanding the Problem Space for Effective Use of a Circular Economy Monitor in Policy Making
No full textInternational audienceThis paper identifies and validates the challenges hindering the integration of circular economy into evidence-based policy making, and proposes an outlook for enhancing the effective use of circular economy monitors. It highlights the limitations of current circular economy monitoring systems, which often fail to transform circular economy information into actionable knowledge for policy makers. Using the echelon Design Science Research approach, which divides projects into manageable ‘echelons’ to tackle complex socio-technical problems, this study focuses on the problem analysis echelon. Through 13 semi-structured interviews with intended users and an extensive literature review, the study identifies and validates five challenges to embedding circular economy in policy-making. These challenges are the delayed benefits of circular economy actions, fragmented policy coordination, the lack of a policy agenda for higher R strategies, the complexity of circular economy implementation, and the gap between theoretical frameworks and practical policy needs. This analysis is grounded in the theory of effective use as our Kernel Theory. To address the ineffective use of circular economy monitors, the study proposes an outlook of design requirements and design principles. This paper contributes to the literature on policy monitoring frameworks and circular economy policymaking by delineating a validated problem space that future researchers can use to improve the effective use of circular economy monitors in policymaking
Exact and Heuristic Methods for Planning and Scheduling Collaborative Manufacturing Systems
No full textPart 2: Collaborative Manufacturing Systems in the Digital EraInternational audienceEmerging challenges in the manufacturing industry, such as supply chain disruptions, political instability, and difficulty in accessing skilled workforce necessitate a more pragmatic approach to survive in this highly intertwined ecosystem. These approaches aim to empower manufacturing companies’ efficiency and agility while providing a certain degree of resilience. Hence, collaboration among stakeholders by sharing manufacturing resources and information is vital. However, despite the advancements in digital platforms for collaborative manufacturing, there is a need for effective planning of shared resources which is computationally intractable. Approaching this challenging problem from a multi-agent perspective brings new opportunities for modeling and solving. In a collaborative manufacturing network, as a multi-agent system, each manufacturing stakeholder, or agent, can pursue their objective, such as minimizing production time, reducing costs, or improving product quality while a coordinator agent monitors and ensures a solution that is best for all agents. This paper proposes systematic and heuristic methods for planning and scheduling collaborative manufacturing resources using a multi-agent modeling paradigm. The efficiency of the developed methods is benchmarked with randomly generated instances that show promising results for the manufacturing industry
Requirements Derived from Digitalization Patterns
No full textPart 3: Fostering Collaborative and Interoperable Digital Models for Digital Twins: MethodsInternational audienceDigital twins became a major research topic throughout diverse domains that started around 2000 in manufacturing. This paper aims at establishing comprehensible digitalization environments by introducing patterns contributing to structuring digitalization efforts. An underlying conceptualization of digitalization environments consisting of digital models, digital-physical interactions, and physical realizations is proposed and used to depict interdependencies among digital and physical building blocks. Digitalization patterns that conceptualize monitoring and controlling techniques are selected based on the application scenario and raise requirements for processing mechanisms as well as for the realization with physical devices. To reflect on the conceptualization and the digitalization patterns, project MEASURE is introduced. Concepts are applied to the emergency exercise context by addressing the application scenario modelling, the interaction between digital-physical world as well as the realization. The discussion derives requirements from the digitalization patterns and reasons on the underlying meta model linking the building blocks of digitalization environments
Lip Recognition Based on Bi-GRU with Multi-Head Self-Attention
No full textPart 1: Deep LearningInternational audienceCurrently, lip recognition technology is a significant research direction in the field of video understanding in computer vision. It aims to recognize the content expressed by the main characters through the dynamic changes of the lips' visual image. With the development of deep learning and enhanced computer performance, lip recognition techniques have evolved to effectively separate the background and target foreground across different scenes, improving uniformity in models and technical routes. However, despite the success of lip recognition with English datasets, the semantic specificity of Chinese words and the scarcity of open-source Chinese lip recognition datasets have generally led to subpar recognition standards.To address these challenges, this paper introduces a novel model that synergizes two-dimensional and three-dimensional networks. Our approach leverages an improved Convolutional 3D (C3D) network to extract spatiotemporal features effectively. Unlike traditional 2D networks, which lack temporal dynamics, and conventional 3D networks that are prone to overfitting due to deep layers, our enhanced C3D network provides a robust foundation for feature extraction. To capture temporal features more efficiently, we integrate the outputs into a Bi-directional Gated Recurrent Unit (Bi-GRU), combined with a Multi-Head Self-Attention mechanism. This fusion allows for better extraction of semantic and syntactic features, overcoming the limitations of normal RNNs in handling long sequences and the non-parallelizable nature of GRU due to sequence dependence.The effectiveness of the proposed model was validated through experiments on our self-compiled Chinese dataset. We compared it with mainstream networks such as ResNet-18, ResNet-34, and the original C3D model lacking the multi-attention mechanism. The analysis of loss function curves and accuracy curves demonstrated that our model achieves significant improvements, effectively addressing the challenges in lip recognition for the Chinese language and setting a new benchmark for performance in this field
Enhancing Predictive Process Monitoring with Conformal Prediction
No full textPart 2: Machine LearningInternational audienceThis paper introduces a framework that integrates Conformal Prediction (CP) with Predictive Process Monitoring (PPM) to enhance prediction accuracy and reliability by producing prediction intervals with a guaranteed coverage rate. The approach followed fills a significant gap in current research as it provides an effective technique for assessing prediction uncertainty, which is vital for making well-informed decisions in various business sectors. Comprehensive experimental research conducted on various datasets demonstrates the framework's ability and effectiveness in providing accurate and reliable predictions of the remaining time required for the completion of a process trace. This work highlights the significance of measuring uncertainty in predictions, providing a substantial contribution to the areas of PPM and CP. It also offers a solid and trustworthy approach for integrating uncertainty quantification into process mining predictive models that contributes to significantly enhanced decision-support
Generating Profiles of News Commentators with Language Models
No full textPart 1: Reinforcement/Natural LanguageInternational audienceUnderstanding the ebb and flow of online conversation has become a core task in a variety of domains. Public policy, public relations, marketing, and a host of other fields concern themselves with extracting, predicting, and reacting to, changes in the topics being discussed by online users, and the disposition these users have with respect to topics of interest. Creating systems that can automate or simplify this process would have an immediate effect on these endeavors. To that end, this contribution proposes a method of leveraging large language models to process corpus’ of online content and comments to generate a set of descriptive profiles describing the hypothetical positions of a commentator engaged with the content. We propose a method of crafting prompts for language models that tasks them with generating these ‘Ideal Profiles’. This method is used to generate profiles based on a corpus of news articles and their associated comments. To evaluate their utility, learned topic models are fit to the article and comment data, as well as manually constructed sets of comparison profiles. The learned topic models are used to evaluate perplexity and coherence metrics between the generated profiles and evaluation corpus’. This paper highlights results that suggest that the profiles generated contain meaningful topics, and that they have coherence with manually constructed profiles
Bruteware: A Novel Family of Cryptoviral Attacks
No full textInternational audienceWe introduce a novel family of cryptoviral attacks designed to disrupt a system for a period of time determined by the attacker. This allows the attacker to impose a specific cost on the target organization or individual. We also unveil two practical cryptographic schemes that can be used to mount such attacks. The proposed schemes exploit the key generation and signing capabilities of the Trusted Platform Module (TPM) to create machine-bound computationally hard problems. These problems are constructed using time-based and memory-hard cryptographic primitives. Victims are then forced to allocate resources to solve them in order to recover their data. By analyzing detection and prevention techniques, this paper also provides guidance on defensive strategies to thwart the presented attacks
Reduce to the MACs - Privacy Friendly Generic Probe Requests
No full textInternational audienceSince the introduction of active discovery in Wi-Fi networks, users can be tracked via their probe requests. Although manufacturers typically try to conceal Media Access Control (MAC) addresses using MAC address randomisation, probe requests still contain Information Elements (IEs) that facilitate device identification. This paper introduces generic probe requests: By removing all unnecessary information from IEs, the requests become indistinguishable from one another, letting single devices disappear in the largest possible anonymity set. Conducting a comprehensive evaluation, we demonstrate that a large IE set contained within undirected probe requests does not necessarily imply fast connection establishment. Furthermore, we show that minimising IEs to nothing but Supported Rates would enable 82.55% of the devices to share the same anonymity set. Our contributions provide a significant advancement in the pursuit of robust privacy solutions for wireless networks, paving the way for more user anonymity and less surveillance in wireless communication ecosystems