IFIP Open Access Digital Library
Not a member yet
22614 research outputs found
Sort by
Machine Learning Models for Electricity Generation Forecasting from a PV Farm
No full textPart 2: Machine LearningInternational audienceAccurate forecasting of the electricity generation from photovoltaic farms plays a significant role in their proper technical and financial management. Reliable forecasts enable management of inertia and frequency response during contingency events and proper planning of the spinning reserve of PV farms. In this work, six machine learning models applying Convolutional Neural Network, Extreme Learning Machine, Random Forest Regression, Gradient Boosted Regression, AdaBoosted Regression, and K-Nearests Neighbors Regression were proposed to forecast electricity generation from a 700 kW photovoltaic farm located in Poland. The models were developed based on four widely available meteorological parameters: ambient air temperature, cloudopacity, and relative humidity, and hour, day, month and year. The comparative performance of the model revealed that the gradient-boosted regression was the most reliable with the determination coefficient R2 = 95.503%, the mean absolute error MAE = 15.003 kWh, and the root mean square error RMSE = 29.975 kWh
Learning-Based Short-Term Energy Consumption Forecasting
No full textPart 2: Machine LearningInternational audienceDevelopment of reliable methods is essential to understand building energy consumption. Traditional statistical models showed drawbacks to express non-linear predictions. The recent artificial intelligence methods are more suitable to study the non-linear correlation between the consumed data, meteorological data, and other features. To address these challenges, we evaluate and compare the performances of ten learning-based models on four energy consumption datasets. The proposed framework includes four preprocessing steps namely, outliers and missing data processing, resampling processing, data normalization and features reduction. The results revealed the importance of the preprocessing steps having a high impact on the forecast performances. In addition, finding results showed that performances drop when resampling the original data values and performances increase when reducing the features by applying Pearson Correlation Coefficient. Based on four evaluation metrics (MAE, MAPE, R-Squared and Pbias), forecast results revealed that the applied models achieved high forecasting performances. Moreover, Machine learning models achieved slightly better performances, especially ensemble models such as ERTR and XGBOOST, outperforming Deep learning models and Hybrid models
Queuing Theoretic Analysis of Dynamic Attribute-Based Access Control Systems
No full textInternational audienceAccess resolution in Attribute-based Access Control (ABAC) is primarily through the enforcement of an ABAC policy. However, incremental user-specific authorizations are also often added to supplement the attribute-based accesses. As this auxiliary list of authorizations grows, enforcement becomes increasingly more inefficient, since both the ABAC policy and the specific authorizations are to be evaluated. Regenerating the ABAC policy from the auxiliary list, on the other hand, requires re-running the computationally expensive policy mining algorithms. Further, access mediation has to be put on hold while policy rebuilding is done, resulting in periods of unavailability of the system. In this paper, we look into the problem of balancing access request resolution, accommodating dynamic authorization updates, and ABAC policy rebuilding. We employ a queuing theoretic approach where the access mediation process is modeled as an M/G/1 queue with vacation. The server is primarily involved in resolving access requests, but occasionally goes on vacation to rebuild the ABAC policy. We study the effects of several parameters like request arrival rate, access resolution time, vacation duration and interval between vacations. Our extensive experiments provide a direction towards efficient implementation of ABAC
Satellite: Effective and Efficient Stack Memory Protection Scheme for Unsafe Programming Languages
No full textInternational audienceMemory unsafe languages are still widely used in a lot of critical software such as operating system kernels and browsers, and thus, memory corruption attacks remain a significant threat. To mitigate the threat, various defense approaches have been proposed such as the Address Space Layout Randomization (ASLR) and Stack Canary. However, adversaries have demonstrated the capability to bypass them, which results in upgraded defense systems. To complement the ASLR, the stack isolation technique that conceals sensitive objects stored in stack memory by relocating them to a “safe region” was introduced. Nonetheless, advanced information disclosure attacks, such as Allocation Oracle, have been employed to discover the location of the safe region. In this work, we introduce Satellite as an effective and efficient approach for safeguarding the stack memory against memory vulnerabilities and information disclosure attacks. The proposed technique guarantees the safety of the return address stored in the safe region, protecting it from vulnerabilities like buffer overflows and information disclosure attacks. To easily support general C/C++ programs, we implemented Satellite in the LLVM compiler framework. To demonstrate the efficiency of Satellite, we applied Satellite to SPEC CPU2006, SPEC CPU2017, and the Nginx web server to assess the effectiveness of the proposed technique. The assessment findings indicate that Satellite incurs an average performance overhead of 0.29
IPEQ: Querying Multi-attribute Records with Inner Product Encryption
No full textInternational audienceWe introduce a secure query processing employing a function hiding inner product encryption (FHIPE) to enable queries on encrypted data that achieve minimal leakage, low latency, and space efficiency. Performing DB operations on encrypted data requires specialized encryption schemes that carefully balance security and performance. For example, fully homomorphic encryption (FHE) is a technique that allows querying an encrypted DB, but FHE ciphertexts are large and can overwhelm machine memory. Querying under FHIPE shows the DB server whether an encrypted record and an encrypted query condition are matched or not using the inner product. However, the query condition and record representation must be carefully considered to ensure that correct results are returned. In this paper, we propose a novel encrypted querying scheme IPEQ with FHIPE. IPEQ can correctly perform multi-attribute Equality, GROUP BY, and JOIN over encrypted data by employing a special encoding to the query condition and record. Our solution also outperforms homomorphically encrypted DB in terms of query execution latency and DB size
Higher-Order Adaptive Dynamical System Modelling of the Role of Epigenetics in Major Depressive Disorder
No full textPart 1: Biomedical/ClassificationInternational audienceThis paper focuses on the modelling of an experiment concerning the emotional response as a result of thinking about happy memories for non-depressed and depressed individuals. Furthermore, it shows the role of epigenetic changes in this experiment by showing that the depletion of the brain-derived neurotrophic factor (BDNF) gene expression, often associated in literature both with depression and memory functions, can lead to a depressed individual and a change of feeling state with happy memory recall. An adaptive dynamical system model was designed for this process, which involves higher-order adaptation in emotion regulation. This computational model was represented as a higher-order adaptive network model and showcases an individual who, at first, does not have a major depressive disorder (MDD), but then a stressful life event happens, leading to epigenetic changes associated with BDNF depletion that make the person develop depression
Semantic Modelling for Representation and Integration of Health Data from Wearable Devices
No full textPart 1: Biomedical/ClassificationInternational audienceThe efficient sharing and integration of health-related data are critical objectives in the rapidly developing field of healthcare technology. In order to achieve this, a seamless exchange of health information across different frameworks is needed, promoting interoperability among health data formats, a key factor in the progression of digital health solutions. In this paper, we present our work in defining and implementing a mapping between the Fast Healthcare Interoperability Resources (FHIR) standard and the Open mHealth (OMH) schema, two well-known and widely used solutions for health-related data modeling. We elaborate on key mapping choices that have been made to conceptually align the two paradigms, and we present examples that demonstrate the applicability of our approach. The contribution to the research community is marked by advancements in interoperability and the facilitation of seamless integration across diverse health data formats. Also, the potential impact is significant, offering benefits to both healthcare professionals and the patients under their care
A Socio-cultural Perspective on Technology for Environmental Sustainability: The Case of Filtering Water Pots (G-filters) in Rajasthan, India
No full textPart 3: General TrackInternational audienceThe design and adoption of socio-culturally appropriate technologies are critical to addressing the environmental sustainability crisis in the Global South. However, implementing technology-based sustainability initiatives has been shown to be problematic, with many such projects failing to achieve their stated objectives or have any meaningful impact at scale. This is because such endeavours are complex, multifaceted, and interdisciplinary and require a comprehensive understanding of the interplay between tradition, technology, politics, climate, and sustainability. Such projects are also likely to be highly socio-culturally contextual and thus require a deep understanding of the specific context and culture within which they will be built and deployed. This paper presents the results of an ongoing research project in Rajasthan, India, where filtering water pots (G-filters) are used to purify waste-laden water. The water pots are made by the traditional Kumhar potter caste and are used extensively for domestic and agricultural use in the local community. Initial results indicate that these pots fulfil both criteria required for impactful technology-based projects in the Global South; i.e., they are technically effective and socio-culturally appropriate for the context in which they are being used. The paper makes various recommendations based on our findings to date and provides details of the next stages of the project
Drawing a Map in the Sand: Locating an Ethics of Care in the ICT-Related Migration Practices of Older Volunteers in the US Southwest
No full textPart 3: General TrackInternational audienceCritical migration studies have highlighted the risks of using digital technologies in the space of migration, underlying its implication for migrants’ privacy rights, government surveillance, and information capitalism. In U.S. Customs and Border Protection’s Tucson sector, stretching across Arizona, one of the busiest and most dangerous borders in the USA, volunteer organizations work to prevent the death of unauthorized migrants undertaking the journey through the Sonoran Desert. Volunteers, typically elderly, study and map migratory trails to provide water where it is likely to be found by migrants. They employ a combination of paper-based and dated, yet sophisticated technologies for gathering essential data to support this potentially life-saving work. In this article, we discuss their information and data practices, and we argue that their approach is an example of the application of an ethics of care to an informational space. Based on interviews and participant observations, we suggest that volunteers’ general refusal to adopt more efficient data practices indicate both a resistance to change from already ingrained practices, as well as an application of caring ethics within the field of migration and information for the pursuit of social justice
Community and Large-Scale Digital Transformation for Poverty Eradication and Economic Growth in Africa: A Rapid Review of Existing Research for the Period 2013–2023
No full textPart 1: Diverse and Inclusive Digital TransformationInternational audienceThis article investigates recent scholarly research on large-scale community and digital transformation (DT) in Africa that relates to SDG 1 (no poverty) and SDG 8 (decent work and economic growth). The study used a rapid review approach to get an exploratory overview of recent scholarly research on the topic. Thirty-seven scholarly papers were included and analysed, departing from a sociotechnical perspective. The findings present a fragmented picture. Most of the studies were quantitative and cross-sectional. A significant number of studies focused on aspects of agriculture, underscoring the importance of agriculture in terms of economic sustainability in Africa. Many studies do not indicate any conceptual foundations that inform the research, or the results presented. The dimensions of DT that were focused on include drivers and disruptions, enablers and constraints, the DT process, and outcomes. The most significant technology related to DT in Africa is mobile technologies. It is also clear from these studies that the development of appropriate technologies cannot be viewed separately from their contexts. Given this, the geographical coverage of Africa in the studies is too limited. The actual changes in the sociotechnical systems during DT are not adequately researched. Studies only focused in a limited way on novel and niche technologies. There is limited evidence of a focus on gender issues, despite their importance. There is a significant scope for qualitative longitudinal studies, and such studies should be accessible to key decision makers and role players