Last 7 days
Bookmark and Share    
Full Text
Peer Reviewed
See detailTowards an Efficient and Coherent Regulatory Framework on Cybersecurity in the EU: The Proposals for a NIS 2.0 Directive and a Cyber Resilience Act
Cole, Mark David UL; Schmitz, Sandra UL

in Applied Cybersecurity & Internet Governance (2022), 1(1),

Cybersecurity regulation in the EU has long been implemented in a piece- meal fashion resulting in a fragmented regulatory landscape. Recent developments triggered the EU to review its approach which has ... [more ▼]

Cybersecurity regulation in the EU has long been implemented in a piece- meal fashion resulting in a fragmented regulatory landscape. Recent developments triggered the EU to review its approach which has not resulted in the envisaged high level of cyber resilience across the Union. The paper addresses the EU’s limited mandate to regulate cybersecurity and outlines how the internal market rationale serves as a basis to harmonise cybersecurity legislation in the EU Member States. In that regard, the recent Proposal for a NIS 2.0 Directive (adopted by the European Parliament in November 2022) and the Proposal for a Cyber Resilience Act (published in September 2022) highlight how the EU seeks to align legislation and reduce complexity between different, often sectoral regulatory approaches to cybersecurity, while at the same time extending regulation in a view to achieve a high level of cybersecurity across the EU. As regards the latter, the paper also outlines how the Cyber Resilience Act will complement the NIS 2.0 Directive in order to close existing regulatory gaps. [less ▲]

Full Text
See detailBridging NLP and LLOD: Humanities Approaches to Semantic Change
Armaselu, Florentina UL

Speeches/Talks (2022)

Natural language processing (NLP) for detecting lexical semantic change and linguistic linked open data (LLOD) are two areas of research that have shown promising results in the latest years. However ... [more ▼]

Natural language processing (NLP) for detecting lexical semantic change and linguistic linked open data (LLOD) are two areas of research that have shown promising results in the latest years. However, their potential of being considered together for analysing and representing semantic change from a humanistic perspective needs further study and development. The talk presents an overview of theoretical aspects, NLP techniques and LLOD formalisms intended to this purpose, and focuses on a project developed as a humanities use case within the COST Action “Nexus Linguarum - European network for Web-centred linguistic data science.” The discussion includes preliminary thoughts on the conception of a system that combines dictionary information with corpus evidence, and provides multilingual diachronic ontologies for humanities research. [less ▲]

Detailed reference viewed: 59 (20 UL)
Full Text
Peer Reviewed
See detailVers des réparations au titre du colonialisme?
Erpelding, Michel UL

in Annuaire français de droit international (2022), 67

Noting the changes that have occurred in recent years in Western official discourses regarding the legacy of colonialism, this article examines the potential legal consequences of these changes in the ... [more ▼]

Noting the changes that have occurred in recent years in Western official discourses regarding the legacy of colonialism, this article examines the potential legal consequences of these changes in the field of international state responsibility. The first part of the article consists of a survey of recent Western state practice. It observes that while some Western states have lately acknowledged the unjustifiable character of certain acts related to colonialism and have taken measures to repare the consequences of these acts, these measures have, with one exception, assumed an ex gratia character. The second part of the article is dedicated to the intertemporal law principle, which constitutes the main argument used to defend this position. Western states invoke this principle to deny the internationally wrongful character of their past colonial policies and rule out any general obligation to make reparation for the damage caused by these policies. However, the principle of intertemporal law has become subject to various reassessments, which challenge the legal postulates inherited from the past. [less ▲]

Detailed reference viewed: 15 (1 UL)
Full Text
Peer Reviewed
See detailSteel-fibre reinforced concrete in compo-site structures as a mean to increase re-sistance and ductility
Zanon, Riccardo; Schäfer, Markus UL; Ruiz, Gonzalo et al

in Stahlbau (2022), 91(12), 801-811

Steel-fibre reinforced concrete is a well-known material used for decades for industrial floorings, shot-crete, or other specific applications. Its use is now spreading in structural applications as a ... [more ▼]

Steel-fibre reinforced concrete is a well-known material used for decades for industrial floorings, shot-crete, or other specific applications. Its use is now spreading in structural applications as a complement or a substitute for conventional bar-reinforced concrete since the normative framework is ready to provide design approaches for several concrete applications. Nevertheless, some adaptations are needed for implementing steel-fibre reinforced concrete in the design approach for steel-concrete composite struc-tures whereby complementary aspects need to be considered. In this field, besides the increase of tensile properties and durability of the concrete member, a very important contribution may be given by the in-crease of the concrete ductility in compression. This property is conferred by the steel fibres which pro-vide a confinement effect increasing the plastic damage that the concrete matrix can absorb. If this prop-erty is widely accepted, its benefits are limited for the design of conventional concrete sections. Con-versely, for composite sections, the possibility to reach higher strains in compression means ensuring full exploitation of the strength capacity of structural steel. Besides reaching an optimization of already in-use cross-sections, this new material combination enhances the use of higher structural steel strengths in composite structures. [less ▲]

Detailed reference viewed: 13 (0 UL)
Full Text
See detailINVESTIGATION OF SURFACE SPUTTERING AND IONIZATION PROCESSES UNDER NON-REACTIVE LIGHT ION IRRADIATION: TOWARDS 4D SIMS IMAGING
Ost, Alexander Dimitri UL

Doctoral thesis (2022)

The progressive trend to miniaturize samples presents a challenge to materials characterization techniques in terms of both lateral resolution and chemical sensitivity. The latest generation of focused ... [more ▼]

The progressive trend to miniaturize samples presents a challenge to materials characterization techniques in terms of both lateral resolution and chemical sensitivity. The latest generation of focused ion beam (FIB) platforms has allowed to advance in a variety of different fields, including nanotechnology, geology, soil, and life sciences. State-of-the-art ultra-high resolution electron microscopy (EM) devices coupled with secondary ion mass spectrometry (SIMS) systems have enabled to perform in-situ morphological and chemical imaging of micro- and even nanosized objects to better understand materials by studying their properties correlatively. However, SIMS images are prone to artefacts induced by the sample topography as the sputtering yield changes with respect to the primary ion beam incidence angle. Knowing the exact sample topography is crucial to understand SIMS images. Moreover, using non-reactive primary ions (Ne+) produced in a gas field ion source (GFIS) allows to image in SIMS with an excellent lateral resolution of < 20 nm, but it comes with a lower ionization probability compared to reactive sources (e.g., Cs+) and due to small probe sizes only a limited number of atoms are sputtered, resulting in low signal statistics. This thesis focused first on taking advantage of high-resolution in-situ EM-SIMS platforms for applications in specific research fields and to go beyond traditional correlative 2D imaging workflows by developing adapted methodologies for 3D surface reconstruction correlated with SIMS (3D + 1). Applying this method to soil microaggregates and sediments allowed not only to enhance their visualization but also to acquire a deeper understanding of materials’ intrinsic transformation processes, in particular the organic carbon sequestration in soil biogeochemistry. To gain knowledge of the influence of the topography on surface sputtering, using model samples the change of the sputtering yield under light ion bombardment (He+, Ne+) for different ranges of incidence angles of the primary ion beam was studied experimentally. This data was compared to Monte Carlo simulation results and fitted with existing sputtering model functions. We showed thus that these models developed and studied for heavier ions (Ar+, Cs+) are also applicable to light ions (He+, Ne+). Additionally, an algorithm used to correct SIMS images with respect to topographical artefacts resulting from local changes of the sputtering yield was presented. Finally, the contribution of oxygen on positive SI yields was studied for non-reactive primary ions (25 keV Ne+) under high primary ion current densities (up to 10^20 ions/(cm2 ∙ s)). It was shown that in order to maximize and maintain a high ionization probability oxygen needs to be provided continuously to the surface. Secondary ion signal enhancement of up to three orders of magnitude were achieved for silicon, opening the doors for SIMS imaging at both highest spatial resolution and high sensitivity. [less ▲]

Full Text
Peer Reviewed
See detailMitochondrial DNA heteroplasmy distinguishes disease manifestation in PINK1/PRKN-linked Parkinson’s disease
Trinh, Joanne; Hicks, Andrew A.; König, Inke R. et al

in Brain (2022)

Biallelic mutations in PINK1/PRKN cause recessive Parkinson’s disease. Given the established role of PINK1/Parkin in regulating mitochondrial dynamics, we explored mitochondrial DNA (mtDNA) integrity and ... [more ▼]

Biallelic mutations in PINK1/PRKN cause recessive Parkinson’s disease. Given the established role of PINK1/Parkin in regulating mitochondrial dynamics, we explored mitochondrial DNA (mtDNA) integrity and inflammation as disease modifiers in carriers of mutations in these genes. MtDNA integrity was investigated in a large collection of biallelic (n = 84) and monoallelic (n = 170) carriers of PINK1/PRKN mutations, idiopathic Parkinson’s disease patients (n = 67) and controls (n = 90). In addition, we studied global gene expression and serum cytokine levels in a subset. Affected and unaffected PINK1/PRKN monoallelic mutation carriers can be distinguished by heteroplasmic mtDNA variant load (AUC = 0.83, CI:0.74-0.93). Biallelic PINK1/PRKN mutation carriers harbor more heteroplasmic mtDNA variants in blood (p = 0.0006, Z = 3.63) compared to monoallelic mutation carriers. This enrichment was confirmed in iPSC-derived (controls, n = 3; biallelic PRKN mutation carriers, n = 4) and postmortem (control, n = 1; biallelic PRKN mutation carrier, n = 1) midbrain neurons. Lastly, the heteroplasmic mtDNA variant load correlated with IL6 levels in PINK1/PRKN mutation carriers (r = 0.57, p = 0.0074). PINK1/PRKN mutations predispose individuals to mtDNA variant accumulation in a dose- and disease-dependent manner. [less ▲]

Detailed reference viewed: 19 (4 UL)
Full Text
Peer Reviewed
See detailStructure of soliton bound states in the parametrically driven and damped nonlinear systems
Bogdan, M.M.; Charkina, Oksana UL

in Low Temperature Physics (2022), 48

Static soliton bound states in nonlinear systems are investigated analytically and numerically in the framework of the parametrically driven and damped nonlinear Schrödinger equation. We find that the ... [more ▼]

Static soliton bound states in nonlinear systems are investigated analytically and numerically in the framework of the parametrically driven and damped nonlinear Schrödinger equation. We find that the ordinary differential equations, which determine bound soliton solutions, can be transformed into the form resembling the Schrödinger-like equations for eigenfunctions with fixed eigenvalues. We assume that a nonlinear part of the equations is close to the reflectionless potential well occurring in the scattering problem associated with the integrable equations. We show that symmetric two-hump soliton solution is quite well described analytically by the three-soliton formula with the fixed soliton parameters, depending on the strength of parametric pumping and the dissipation constant. [less ▲]

Detailed reference viewed: 8 (0 UL)
See detailTransnationale Begegnungen in Soldatenbriefen - Beispiel WARLUX – Briefe von Luxemburgern in der deutschen Wehrmacht
Janz, Nina UL

Speeches/Talks (2022)

Kriege sind Räume unterschiedlicher Begegnungen und Handlungsfelder: sei es durch Menschen, die durch Flucht und Migration, durch Deportation oder Gefangenschaft gezwungen sind, ihre Länder zu verlassen ... [more ▼]

Kriege sind Räume unterschiedlicher Begegnungen und Handlungsfelder: sei es durch Menschen, die durch Flucht und Migration, durch Deportation oder Gefangenschaft gezwungen sind, ihre Länder zu verlassen, oder durch die Bewegungen von Soldaten. Während des Zweiten Weltkriegs waren beispielsweise über 50 Millionen Soldaten auf dem Marsch. Sie verließen ihre Familien und ihre Heimat auf Befehl oder freiwillig, sie wurden in fremden Ländern eingesetzt, sie kämpften mit verschiedenen anderen Nationalitäten auf derselben Seite, und sie interagierten mit der einheimischen Bevölkerung, sei es freundlich oder feindlich. Transnationale Begegnungen und die Verflechtung von Wegen und Leben waren im Krieg an der Tagesordnung. In der Wehrmacht zum Beispiel kämpften viele verschiedene Nationen gemeinsam auf derselben Seite, sei es durch Pflichtdienst, freiwilligen Dienst oder Zwang. Die Wehrmacht war tatsächlich eine transnationale Armee, auch wenn dies unter dem Begriff "Deutschtum" versteckt wurde. Der militärische Raum und die transnationalen Interaktionen stehen im Mittelpunkt meiner aktuellen Forschung. Basierend auf dem WARLUX-Projekt "Luxemburger und die Kriegserfahrung" suche ich nach Möglichkeiten, diese transnationalen Begegnungen zu identifizieren. Das Projekt verfolgt einen biographischen Ansatz und konzentriert sich auf die einzelnen Akteure: die Soldaten und ihre transnationalen Kriegserfahrungen. Anhand von Ego-Dokumenten, insbesondere Feldpostbriefen, untersucht das Projekt die individuellen Bezüge zu diesen Begegnungen. In den Briefen schreiben die Luxemburger oft von "wir" als deutsche Soldaten, dann aber auch von "sie", den Deutschen. Sie schreiben oft von Kontakten mit "ihren" Leuten oder Brüdern, wenn sie auf andere transnationale Soldaten trafen, wie z.B. auf wehrpflichtige Elsässer. In anderen Momenten identifizierten sie sich wieder als "deutsche Soldaten", etwa wenn es darum ging, die anderen, d.h. die Alliierten und die Rote Armee, zurückzuschlagen, um selbst zu überleben. Kann mit diesen Methoden und auf der Basis der Feldpostbriefe eine "Transnationalität" hergestellt werden? Das Projekt wird mit diesen Methoden durch HTR und digitale Transkriptionen dank einer Textanalyse und Distant und Close Reading experimentieren und die Bedeutung in Bezug auf diese Personengruppe und die Quellen überprüfen. [less ▲]

Detailed reference viewed: 15 (1 UL)
See detailAnalysis of End-of-Life Process Data to enable Design Knowledge for Circularity
Mangers, Jeff UL

Doctoral thesis (2022)

The concept of Circular Economy (CE) is gaining increasing attention as an indispensable renewal of linear economy without neglecting sustainable development goals. Closing resource loops and keeping ... [more ▼]

The concept of Circular Economy (CE) is gaining increasing attention as an indispensable renewal of linear economy without neglecting sustainable development goals. Closing resource loops and keeping resources in the system at the highest level of use for as long as possible are cited as the main goals of CE. However, due to missing information exchange, the lack of consistency between the existing End-of-Life (EOL) infrastructure and the respective product designs hinders a successful circularity of resources. This research provides a modular method to collect, process, and apply EOL process data to provide the Beginning-of-Life (BOL) with important EOL-knowledge through a CE adapted product design assessment. EOL-data is collected using a Circular Value Stream Mapping (CVSM), EOL-information is processed using a digital state flow representation, and EOL-knowledge is applied by providing a graphical user interface for designers. The method is verified by a simulation model that serves as a decision-support tool for product designers in the context of a PET bottle case study in Luxembourg. The goal is to anticipate a circular flow of resources by reflectively aligning product design with the relevant EOL infrastructure. Within the linear economy, the focus has been on improving production processes while neglecting what happens to a product after its use. The developed method makes it possible to consider not only the requirements of users but also the actual end users, the EOL process chains, when designing products. [less ▲]

Detailed reference viewed: 18 (1 UL)
Full Text
Peer Reviewed
See detailRate Splitting Multiple Access for Cognitive Radio GEO-LEO Co-Existing Satellite Networks
Khan, Wali Ullah UL; Ali, Zain; Lagunas, Eva et al

Scientific Conference (2022, December 07)

Low Earth orbit (LEO) satellite communication has drawn particular attention recently due to its high data rate services and low round-trip latency. It is low-cost to launch and can provide global ... [more ▼]

Low Earth orbit (LEO) satellite communication has drawn particular attention recently due to its high data rate services and low round-trip latency. It is low-cost to launch and can provide global coverage. However, the spectrum scarcity might be one of the critical challenges in the growth of LEO satellites, impacting severe restrictions on the development of ground-space integrated networks. To address this issue, we propose rate splitting multiple access (RSMA) for cognitive radio (CR) enabled nongeostationary orbit (GEO)-LEO coexisting satellite network. In particular, this work aims to maximize the system's sum rate by simultaneously optimizing the power allocation and subcarrier beam assignment of LEO satellite communication while restricting the interference temperature to GEO satellite users. The problem of sum rate maximization is formulated as non-convex and a Global optimal solution is challenging to obtain. Therefore, we first employ the successive convex approximation technique to reduce the complexity and make the problem more tractable. Then for the power allocation, we exploit Karush–Kuhn–Tucker (KKT) condition and adopt an efficient algorithm based on the greedy approach for subcarrier beam assignment. We also propose two suboptimal schemes with fixed power allocation and random subcarrier beam assignment. [less ▲]

Detailed reference viewed: 20 (1 UL)
Full Text
Peer Reviewed
See detailCODEBERT-NT: code naturalness via CodeBERT
Khanfir, Ahmed UL; Jimenez, Matthieu UL; Papadakis, Mike UL et al

in 22nd IEEE International Conference on Software Quality, Reliability and Security (QRS'22) (2022, December 05)

Much of recent software-engineering research has investigated the naturalness of code, the fact that code, in small code snippets, is repetitive and can be predicted using statistical language models like ... [more ▼]

Much of recent software-engineering research has investigated the naturalness of code, the fact that code, in small code snippets, is repetitive and can be predicted using statistical language models like n-gram. Although powerful, training such models on large code corpus can be tedious, time consuming and sensitive to code patterns (and practices) encountered during training. Consequently, these models are often trained on a small corpus and thus only estimate the language naturalness relative to a specific style of programming or type of project. To overcome these issues, we investigate the use of pre-trained generative language models to infer code naturalness. Pre-trained models are often built on big data, are easy to use in an out-of-the-box way and include powerful learning associations mechanisms. Our key idea is to quantify code naturalness through its predictability, by using state-of-the-art generative pre-trained language models. Thus, we suggest to infer naturalness by masking (omitting) code tokens, one at a time, of code-sequences, and checking the models’ability to predict them. We explore three different predictability metrics; a) measuring the number of exact matches of the predictions, b) computing the embedding similarity between the original and predicted code, i.e., similarity at the vector space, and c) computing the confidence of the model when doing the token completion task regardless of the outcome. We implement this workflow, named CODEBERT-NT, and evaluate its capability to prioritize buggy lines over non-buggy ones when ranking code based on its naturalness. Our results, on 2,510 buggy versions of 40 projects from the SmartShark dataset, show that CODEBERT-NT outperforms both, random-uniform and complexity-based ranking techniques, and yields comparable results to the n-gram models. [less ▲]

Detailed reference viewed: 26 (1 UL)
Full Text
Peer Reviewed
See detailGEO Payload Power Minimization: Joint Precoding and Beam Hopping Design
Ha, Vu Nguyen UL; Nguyen, Ti Ti; Lagunas, Eva UL et al

in Proceedings of IEEE Global Communications Conference GLOBECOM 2022 (2022, December 05)

This paper aims to determine linear precoding (LP) vectors, beam hopping (BH), and discrete DVB-S2X transmission rates jointly for the GEO satellite communication systems to minimize the payload power ... [more ▼]

This paper aims to determine linear precoding (LP) vectors, beam hopping (BH), and discrete DVB-S2X transmission rates jointly for the GEO satellite communication systems to minimize the payload power consumption and satisfy ground users’ demands within a time window. Regarding constraint on the maximum number of illuminated beams per time slot, the technical requirement is formulated as a sparse optimization problem in which the hardware-related beam illumination energy is modeled in a sparsity form of the LP vectors. To cope with this problem, the compressed sensing method is employed to transform the sparsity parts into the quadratic form of precoders. Then, an iterative window-based algorithm is developed to update the LP vectors sequentially to an efficient solution. Additionally, two other two-phase frameworks are also proposed for comparison purposes. In the first phase, these methods aim to determine the MODCOD transmission schemes for users to meet their demands by using a heuristic approach or DNN tool. In the second phase, the LP vectors of each time slot will be optimized separately based on the determined MODCOD schemes. [less ▲]

Detailed reference viewed: 21 (2 UL)
Full Text
Peer Reviewed
See detailHarmonizing sensitive data exchange and double-spending prevention through blockchain and digital wallets: The case of e-prescription management
Schlatt, Vincent; Sedlmeir, Johannes UL; Traue, Janina et al

in Distributed Ledger Technologies: Research and Practice (2022)

The digital transformation of the medical sector requires solutions that are convenient and efficient for all stakeholders while protecting patients’ sensitive data. One example that has already attracted ... [more ▼]

The digital transformation of the medical sector requires solutions that are convenient and efficient for all stakeholders while protecting patients’ sensitive data. One example that has already attracted design-oriented research are medical prescriptions. However, current implementations of electronic prescription management systems typically create centralized data silos, leaving user data vulnerable to cybersecurity incidents and impeding interoperability. Research has also proposed decentralized solutions based on blockchain technology, but privacy-related challenges have often been ignored. We conduct design science research to develop and implement a system for the exchange of electronic prescriptions that builds on two blockchains and a digital wallet app. Our solution combines the bilateral, verifiable, and privacy-focused exchange of information between doctors, patients, and pharmacies through verifiable credentials with a token-based, anonymized double-spending check. Our qualitative and quantitative evaluations as well as a security analysis suggest that this architecture can improve existing approaches to electronic prescription management by offering patients control over their data by design, a high level of security, sufficient performance and scalability, and interoperability with emerging digital identity management solutions for users, businesses, and institutions. We also derive principles on how to design decentralized, privacy-oriented information systems that require both the exchange of sensitive information and double-usage protection. [less ▲]

Full Text
Peer Reviewed
See detailDynamic Beam-Layout Design for MEO High Throughput Satellite Systems
Chaker, Haythem UL; Chougrani, Houcine UL; Alves Martins, Wallace UL et al

Scientific Conference (2022, December 04)

We propose a traffic-oriented beam-layout optimization framework for a medium Earth orbit (MEO) high throughput satellite (HTS) system. The designed beam-layout plans have the objective of dynamic traffic ... [more ▼]

We propose a traffic-oriented beam-layout optimization framework for a medium Earth orbit (MEO) high throughput satellite (HTS) system. The designed beam-layout plans have the objective of dynamic traffic load balancing with minimal radio resource management. We model and evaluate the high directivity and the high reconfigurability capabilities of next-generation HTS systems that are equipped with a processing power on board. The resulting payload flexibility is compared numerically against the on-ground counterpart. Adequate key performance indicators, such as the Jain’s fairness index, the load distribution gap and optimization convergence time, are used to benchmark the system against state-of-the-art solutions. Results show that a dynamic HTS in MEO is capable of efficiently attaining high quality load balancing performance under realistic spatiotemporally varying traffic demands. [less ▲]

Detailed reference viewed: 91 (15 UL)
Full Text
Peer Reviewed
See detailCurvature in Biological Systems: Its quantification, Emergence and Implications Across the Scales
Schamberger, Barbara; Roschger, Andreas; Sengupta, Anupam UL et al

in Advanced Materials (2022)

Surface curvature both emerges from, and influences the behavior of, living objects at length scales ranging from cell membranes to single cells to tissues and organs. The relevance of surface curvature ... [more ▼]

Surface curvature both emerges from, and influences the behavior of, living objects at length scales ranging from cell membranes to single cells to tissues and organs. The relevance of surface curvature in biology has been supported by numerous recent experimental and theoretical investigations in recent years. In this review, we first give a brief introduction to the key ideas of surface curvature in the context of biological systems and discuss the challenges that arise when measuring surface curvature. Giving an overview of the emergence of curvature in biological systems, its significance at different length scales becomes apparent. On the other hand, summarizing current findings also shows that both single cells and entire cell sheets, tissues or organisms respond to curvature by modulating their shape and their migration behavior. Finally, we address the interplay between the distribution of morphogens or micro-organisms and the emergence of curvature across length scales with examples demonstrating these key mechanistic principles of morphogenesis. Overall, this review highlights that curved interfaces are not merely a passive by-product of the chemical, biological and mechanical processes but that curvature acts also as a signal that co-determines these processes. [less ▲]

Detailed reference viewed: 65 (0 UL)
Full Text
See detailLes Italien-ne-s dans le Luxembourg des migrations (1ère partie)
Scuto, Denis UL

Article for general public (2022)

Detailed reference viewed: 27 (0 UL)
Full Text
Peer Reviewed
See detailFaster and Cheaper Energy Demand Forecasting at Scale
Bernier, Fabien UL; Jimenez, Matthieu UL; Cordy, Maxime UL et al

in Has it Trained Yet? Workshop at the Conference on Neural Information Processing Systems (2022, December 02)

Energy demand forecasting is one of the most challenging tasks for grids operators. Many approaches have been suggested over the years to tackle it. Yet, those still remain too expensive to train in terms ... [more ▼]

Energy demand forecasting is one of the most challenging tasks for grids operators. Many approaches have been suggested over the years to tackle it. Yet, those still remain too expensive to train in terms of both time and computational resources, hindering their adoption as customers behaviors are continuously evolving. We introduce Transplit, a new lightweight transformer-based model, which significantly decreases this cost by exploiting the seasonality property and learning typical days of power demand. We show that Transplit can be run efficiently on CPU and is several hundred times faster than state-of-the-art predictive models, while performing as well. [less ▲]

Detailed reference viewed: 29 (3 UL)