HAL Portal IMT Nord Europe
Not a member yet
6687 research outputs found
Sort by
Exploiter les motifs graduels pour améliorer l’explicabilité des modèles
No full textInternational audienc
Auction-Based Approach for Pickup and Delivery Problem with Time Windows
No full textInternational audienc
REACT 2025: the Third Multiple Appropriate Facial Reaction Generation Challenge
No full textInternational audienceIn dyadic interactions, a broad spectrum of human facial reactions might be appropriate for responding to each human speaker behaviour. Following the successful organisation of the REACT 2023 and REACT 2024 challenges, we are proposing the REACT 2025 challenge encouraging the development and benchmarking of Machine Learning (ML) models that can be used to generate multiple appropriate, diverse, realistic and synchronised human-style facial reactions expressed by human listeners in response to an input stimulus (i.e., audio-visual behaviours expressed by their corresponding speakers). As a key of the challenge, we provide challenge participants with the first natural and large-scale multimodal Multiple Appropriate Facial Reaction Generation (MAFRG) dataset (called MARS) recording 136 human-human dyadic interactions containing a total of 2856 interaction sessions covering</div
Preventing crack propagation for self-healing potential in PVA-modified CSH composites
No full textInternational audienc
Study of rubber coating effect on the stability and fracture resistance of railway ballast
No full textInternational audienc
Investigating the impact of polyvinyl alcohol on mechanical and thermal properties of expansive soil under freeze–thaw cycles
No full textInternational audienc
Process-aware Plug-Assisted Thermoforming: Integrating In-Operando Thermal and Strain Measurements for Enhanced Simulation Accuracy
No full textAcknowledgements: The authors acknowledge IMT Nord Europe and the Hauts-de-France Region for supporting this research during the PhD contract of Dr. Dastidar (2018–2022).International audienceProcess-aware Plug-Assisted Thermoforming: Integrating In-Operando Thermal and Strain Measurements for Enhanced Simulation Accurac
Measure estimation on a manifold explored by a diffusion process
No full textInternational audienceFrom the observation of a diffusion path on a compact connected -dimensional manifold without boundary, we consider the problem of estimating the stationary measure of the process. Wang and Zhu (2023) showed that for the Wasserstein metric and for , the convergence rate of is attained by the occupation measure of the path when is a Langevin diffusion. We extend their result in several directions. First, we show that the rate of convergence holds for a large class of diffusion paths, whose generators are uniformly elliptic. Second, the regularity of the density of the stationary measure with respect to the volume measure of can be leveraged to obtain faster estimators: when belongs to a Sobolev space of order , smoothing the occupation measure by convolution with a kernel yields an estimator whose rate of convergence is of order . We further show that this rate is the minimax rate of estimation for this problem
On Disaster-Resilient Hybrid Service Function Chains
No full textInternational audienceNetwork Function Virtualization (NFV), offers a flexible solution through the orchestration of Virtual Network Functions (VNFs) into a Service Function Chain (SFC), facilitating the seamless transfer of user data between customers and edge servers/clouds. The distinct requirements of forward and backward traffic necessitate a Hybrid SFC (HSFC) approach. However, the resilience of these NFV and SFC deployments against disasters, whether natural or due to hardware malfunctions, presents a significant challenge, often resulting in service disruptions or degraded performance within disaster zones (DZ). To address these challenges, we propose the Resilient Hybrid Service Function Chain Resource Optimization (R-HSFC-RO) strategy, aimed at minimizing network resource consumption while ensuring service resilience. This strategy accounts for bandwidth allocation, computing resource requirements for executing VNFs, installation costs, and latency considerations. We propose a novel heuristic approach for a scalable solution, leveraging the constrained shortest path algorithm in a multi-partite graph framework. Simulation results affirm the R-HSFC-RO strategy's effectiveness in enhancing resource and cost efficiency, yielding gains of up to 30% in overall cost and up to 50% in CPU and VNF costs, thereby improving the resilience and operational efficiency of HSFC deployments against potential disruptions