HAL-CentraleSupelec
Not a member yet
77624 research outputs found
Sort by
Machine Learning in Legal Decision-Making: Analysis of Judicial and Algorithmic Reasoning in Road Homicide Cases
No full textInternational audienceCrime scene analysis requires the evaluation of multiple factors to determine a suspect's guilt, a process that can be lengthy and costly. The integration of Artificial Intelligence into the judicial system is emerging as an opportunity to improve the efficiency of investigations and legal decision-making. In this study, we propose a Machine Learning-based methodology to support the assessment of road homicide cases under Italian law. Our approach employs a Large Language Model to extract 51 features from crime scene descriptions automatically. Four Machine Learning models are then analyzed: Random Forest, Gradient Boosting Machine, Decision Tree, and Logistic Regression. We evaluated the performance of these models on a dataset of 100 road homicide court rulings in Italy, achieving 95% accuracy in crime classification. The validation was conducted by comparing the model outputs with a legal ranking established by four legal experts, allowing us to verify the consistency of the algorithmic predictions with human legal reasoning. The results indicate that the Gradient Boosting Machine shows the highest correlation with legal evaluations (ρ = 0.857, τ = 0.714, p-values = 0.05), while Logistic Regression performed the worst. This study highlights the potential of Artificial Intelligence in legal decision support, emphasizing the need to ensure transparency and bias mitigation to comply with European Union Regulations while maintaining human judgment as the central authority in legal proceedings
Bias in genome-wide association test statistics due to omitted interactions
No full textInternational audienceOver the past two decades, genome-wide association studies (GWAS) enabled the discovery of thousands of variants associated with many complex human traits. However, conventional GWAS are still widely performed with linear models with the assumption that the genetic effects are predominantly additive. In this work, we investigate the test statistic behavior when linear models are used to obtain significant genotype-phenotype associations without accounting for epistasis. We first algebraically derive mean and variance shift in the null statistic due to the omitted interaction term, and define the boundary between conservative (i.e., deflated statistic tail) and anti-conservative (i.e., inflated statistic tail) regimes for the common GWAS significance threshold. We then perform phenotype simulation analyses using the Estonian Biobank genotypes and validate the mathematical model. We demonstrate that the anticonservative regime is plausible under realistic parameter settings and models omitting interaction terms can produce spurious significance. Our findings suggest caution when interpreting statistically significant signals reported in the literature based on linear models, especially for large-scale GWAS
A programming language combining quantum and classical control
No full textThe two main notions of control in quantum programming languages are often referred to as "quantum" control and "classical" control. With the latter, the control flow is based on classical information, potentially resulting from a quantum measurement, and this paradigm is well-suited to mixed state quantum computation. Whereas with quantum control, we are primarily focused on pure quantum computation and there the "control" is based on superposition. The two paradigms have not mixed well traditionally and they are almost always treated separately. In this work, we show that the paradigms may be combined within the same system. The key ingredients for achieving this are: (1) syntactically: a modality for incorporating pure quantum types into a mixed state quantum type system; (2) operationally: an adaptation of the notion of "quantum configuration" from quantum lambda-calculi, where the quantum data is replaced with pure quantum primitives; (3) denotationally: suitable (sub)categories of Hilbert spaces, for pure computation and von Neumann algebras, for mixed state computation in the Heisenberg picture of quantum mechanics
Electric mobility investment in developing countries: Emerging patterns from cross-country analysis
No full textInternational audienceOn paper, many developing countries have made pledges to decarbonise and reduce GHG emissions. Nevertheless,decarbonisation is barely happening in many of them. Among other reasons, transport electrificationwith electric mobility, which is pivotal in the decarbonisation strategy of many developing countries, is beset byinvestment challenges on the demand and supply sides. On the demand side, the lingering ‘where will the moneycome from’ challenge remains critical because governments are financially constrained. On the supply side,investors remain uncertain about which country to invest in for electric mobility. This paper contributes to theacademic and policy debate from the transport and power sector coupling context. We apply our conceptualframework to analyse some developing countries with wholesale power markets and wholesale and retail powermarkets. Then, we conduct cross-country analysis of fifteen countries to assess the possibility of emerging patternsof electric mobility investment solutions in these countries. We argue that regional or continental patternsmay be emerging among some countries. We found that Latin American countries (Chile, Colombia, Argentina)appear to follow a pattern of investment in public transport electrification with electric buses, although thecountries are at different stages of development. Romania, Poland, and Türkiye appear to follow an investmentpattern in private electric vehicles. We found a pattern with low-cost two and three-wheelers in Central Americancountries (Nicaragua, Guatemala, and El Salvador). We recommend policies on electric mobility investmentissues in developing countries
Revisiting PQ WireGuard: A Comprehensive Security Analysis With a New Design Using Reinforced KEMs
No full textInternational audienceWireGuard is a VPN based on the Noise protocol, known for its high performance, small code base, and unique security features. Recently, Hülsing et al. (IEEE S&P'21) presented post-quantum (PQ) WireGuard, replacing the Diffie-Hellman (DH) key exchange underlying the Noise protocol with key-encapsulation mechanisms (KEMs). Since WireGuard requires the handshake message to fit in one UDP packet of size roughly 1200 B, they combined Classic McEliece and a modified variant of Saber. However, as Classic McEliece public keys are notoriously large, this comes at the cost of severely increasing the server's memory requirement. This hinders deployment, especially in environments with constraints on memory (allocation), such as a kernel-level implementations.In this work, we revisit PQ WireGuard and improve it on three fronts: design, (computational) security, and efficiency. As KEMs are semantically, but not syntactically, the same as DH key exchange, there are many (in hindsight) ad-hoc design choices being made, further amplified by the recent finding on the binding issues with PQ KEMs (Cremers et al., CCS'24). We redesign PQ WireGuard addressing these issues, and prove it secure in a new computational model by fixing and capturing new security features that were not modeled by Hülsing et al. We further propose 'reinforced KEM' (RKEM) as a natural building block for key exchange protocols, enabling a PQ WireGuard construction where the server no longer needs to store Classical McEliece keys, reducing public key memory by 190 to 390×. In essence, we construct a RKEM named 'Rebar' to compress two ML-KEM-like ciphertexts which may be of an independent interest
Smart Digital Environments for Monitoring Precision Medical Interventions and Wearable Observation and Assistance
No full textInternational audienceVarious recurring medical events encourage innovative patient well-being through connected health strategies based on an elegant digital environment that prioritizes safety, comfort, and beneficial outcomes for both patients and medical staff. This narrative review article aims to investigate and highlight the potential of advanced, reliable, high-precision, and secure medical observation and intervention missions. These involve a smart digital environment integrating smart materials combined with smart digital monitoring. These medical implications concern robotic surgery and drug delivery through image-assisted implantation, as well as wearable observation and assistive tools. The former requires high-precision motion and positioning strategies, while the latter enables sensing, diagnosis, monitoring, and central task assistance. Both advocate minimally invasive or noninvasive procedures and precise supervision through autonomously controlled processes with staff participation. The article analyzes the requirements and evolution of medical interventions, robotic actuation technologies for positioning actuated and self-moving instances, monitoring of image-assisted robotic procedures using digital twins and augmented digital tools, and wearable medical detection and assistance devices. A discussion including future research perspectives and conclusions complete the article. The different themes addressed in the proposed paper, although self-sufficient, are supported by examples of the literature, allowing a deeper understanding
A Correct-by-Construction Checker for Validation of Railway Data
No full textInternational audienceThe objective of our work is the checking of requirements on data describing railway networks. We expose the design, the formal specification and an implementation of a program that checks such constraints, with a formal proof of correctness. That program is compiled into an executable code, linked together with some additional, non-verified code for input/output. It is experimented on significantly complex examples, showing that the efficiency of the checker is competitive with non-formally verified related tools
Spectral properties of magnetic fields on sub-Riemannian contact manifolds
No full textMotivated by some recent studies of the magnetic Laplacian on Riemannian manifolds, we focus on the first eigenvalue of the magnetic horizontal Laplacian on contact manifolds. We characterize conditions for positive spectral shift, and provide some sharp upper bounds. In the Riemannian setting, a genus 1 assumption is known to force the underlying metric to be flat when equality holds in the sharp upper bounds. Interestingly, we show that the equivalent topological condition in the three--dimensional contact setting consists of having first Betti number equal to 2. In this case, equality in our upper bounds implies that the structure is that of a Heisenberg left--invariant nilmanifold. We conclude by showing that, in some specific three--dimensional contact settings, the knowledge of the first eigenvalue of the magnetic Laplacian uniquely determines the manifold Chern class, fully determining the topology of the underlying manifold
Residual stresses and strains during laser assisted tape placement of thermoplastic composite: Multi-physical modelling and experimental validation
No full textInternational audienceThe prediction of residual stresses and strains during laser-assisted tape placement of thermoplastic composite has been investigated by numerous studies in the literature. They are however rarely validated compared to experimental results and based on simplifications of the material thermomechanical behaviour. This study proposes a multi-physical model to address the heat transfer, crystallization kinetics, the induced thermal and crystallization strains as well as the mechanical behaviour of the material during processing. The thermal and crystallization models were already validated in a previous study, and the mechanical behaviour is described with an incremental linear elastic constitutive law and the Classical Lamination Theory. The thermomechanical properties are based on values from the literature and the supplied datasheet. Without any fine tuning of the model, predicted curvatures of cross ply laminates are well described, as well as their evolution with the temperature after manufacturing and during annealing. The measured and calculated curvatures are found in excellent agreement with errors comprised between 2% and 13%. A sensitivity analysis demonstrates that the developed model is more able to correctly reproduce the experimentally observed material behaviour compared to simplified approaches found in the literature
A Proximal Algorithm for Joint Blood Flow Computation and Tissue Motion Compensation in Doppler Ultrafast Ultrasound Imaging
No full textInternational audienceAccurate tissue-clutter rejection and blood flow estimation remain challenging in ultrasound imaging. Traditionally, this estimation is performed by assuming static tissues. Only a few preprocessing techniques attempt to deal with the more realistic but challenging scenario where the tissues are moving. This paper tackles this scenario and presents a novel method for computing blood flow from moving tissues in ultrafast ultrasound imaging. The proposed computational ultrasound imaging method solves a global inverse problem that jointly computes blood flow, tissues, and their motions. The resulting cost function incorporates each component specificity using appropriate regularizations and is fully convex. The cost function is minimized using an alternating proximal-forward-backward algorithm with convergence guarantees. Furthermore, the proposed method is integrated into a multi-resolution scheme for large motions. The experiments show that the proposed method accurately compensates tissue motions, improving the precision of blood flow computation compared to previous methods. Experiments on in vivo images demonstrate the effectiveness of the proposed method in realistic scenarios with large motions.</div