HAL-MINES ParisTech
Not a member yet
27348 research outputs found
Sort by
Trajectoires de socialisation organisationnelle des NEET : le cas des Compagnons du Devoir et du Tour de France
No full textInternational audienc
High-Resolution Range Profile Classifiers Require Aspect-Angle Awareness
No full textWe revisit High-Resolution Range Profile (HRRP) classification with aspect-angle conditioning. While prior work often assumes that aspect-angle information is incomplete during training or unavailable at inference, we study a setting where angles are available for all training samples and explicitly provided to the classifier. Using three datasets and a broad range of conditioning strategies and model architectures, we show that both single-profile and sequential classifiers benefit consistently from aspect-angle awareness, with an average accuracy gain of about 7% and improvements of up to 10%, depending on the model and dataset. In practice, aspect angles are not directly measured and must be estimated. We show that a causal Kalman filter can estimate them online with a median error of 5°, and that training and inference with estimated angles preserves most of the gains, supporting the proposed approach in realistic conditions.</div
Network-based analysis of genome-wide biobank data boosts discovery of genetic associations in psoriasis
No full textPsoriasis is a common autoimmune disease with a strong genetic component. Targeting the IL-23/IL-17 pathways for the treatment of moderate-severe psoriasis has proven successful, which makes it a good benchmark for other drug discovery approaches. Genome Wide Association Studies (GWAS) identify genetic loci associated with a phenotype (e.g. disease). However, using identified loci to understand the disease mechanisms remains challenging. In response, gene networks methods that consider gene-gene relationships have been developed. In a previous study, we showed that the combination of network methods improved the power, reproducibility, and interpretability of those approaches on a small cohort of breast cancer patients. The goal of the present study is to benchmark the performance of the same combination of network methods on a psoriasis 1 .</div
Deep Learning Bridges Histology and Transcriptomics to Predict Molecular Subtypes and Outcomes in Muscle-Invasive Bladder Cancer
No full textAbstract Muscle-Invasive Bladder Cancer (MIBC) is a heterogeneous disease with distinct molecular subtypes influencing prognosis and therapeutic response. However, molecular profiling through RNA sequencing remains costly, time-consuming and complicated by intratumoral heterogeneity. We developed a Deep Learning (DL) approach to infer molecular subtypes from routine histopathological slides and to evaluate its prognostic value in patients treated with neoadjuvant chemotherapy (NAC). We developed an DL-model predicting the expression of 848 subtype-associated genes from histological images of transurethral resection of bladder tumor, enabling spatial molecular subtyping at tile level. The model was trained on 297 NAC-treated patients from the VESPER clinical trial and evaluated on three independent cohorts (COBLAnCE, n=224; Saint-Louis, n=30 and TCGA, n=315), covering diverse staining protocols and scanner types. Spatial transcriptomics from six VESPER patients confirmed the spatial consistency of the inferred expression profiles. Our approach achieved a ROC AUC of 0.94 for molecular subtype prediction, with 95% of genes significantly predicted, demonstrating its ability to capture transcriptomic dysregulations from histological morphology. Predicted expression maps revealed spatially coherent patterns and intratumoral molecular heterogeneity. Importantly, tumors predicted with basal/squamous features (pure or mixed), were associated with significantly worse progression-free and overall survival after NAC (log-rank p=0.014 and 0.037, respectively). This DL-based framework enables accurate and spatially resolved inference of gene expression and molecular subtypes in MIBC without sequencing. These findings could improve patient stratification in clinical practice and support the design of more targeted clinical trials. Further validation in larger cohorts is needed before routine clinical implementation
Modelling of CO₂ injection in Opalinus Clay: a benchmark study on hydraulic-mechanical and hydraulic-chemical effects using different modelling approaches
No full textInternational audienceThe CO₂ Long-term Periodic Injection Experiment (CO₂LPIE) at the Mont Terri rock laboratory is a mid-scale (meter-scale) experiment that bridges the gap between full-scale in situ experiments and laboratory-scale tests. The experiment involves the continuous injection of CO₂ dissolved in a fluid (artificial pore water) into the sandy facies of the Opalinus Clay formation, with periodic variations in injection pressure over several years. By combining comprehensive rock characterization, monitoring of relevant parameters, and numerical modeling, this experiment investigates the complex interactions between CO₂ injection and the resulting hydraulic, mechanical, and chemical effects in Opalinus Clay (Sciandra et al., 2022a,b). As part of this project, a modeling working group composed of five teams is using different software packages—CODE_BRIGHT (Olivella et al., 1996), HYTEC (Van der Lee et al., 2003), OpenGeoSys (OGS 6; Bilke et al., 2019), and PFLOTRAN (Hammond et al., 2014)—to simulate the in situ experiment, with varying focus on coupled hydraulic-mechanical (HM) and hydraulic-chemical (HC) processes. This ongoing study presents a work-in-progress benchmarking effort on CO₂ injection in Opalinus Clay, in which various modeling approaches and setups are explored to assess their impact on the evolution of key process variables
Vapor–Liquid Equilibrium of the Hydrogen Sulfide (H2S) – Benzene (C6H6) Binary System: Experimental and Modeling Study
No full textInternational audienceThe phase behavior of the hydrogen sulfide (H₂S) - benzene (C₆H₆) binary system is critical for optimizing gas sweetening, aromatic solvent recovery, and high-pressure reservoir in the petroleum industry, while ensuring environmental compliance. This study presents new isothermal vapor-liquid equilibrium (VLE) measurements for the H₂S - C₆H₆ system at 278.21 K, 298.36 K, 323.38 K, and 343.39 K, covering pressures up to 4.5 MPa. The experimental data were obtained using a static-analytic method with two magnetic capillary samplers (ROLSI®), enabling precise sampling and analysis of both liquid and vapor phases via gas chromatography. The measurements have uncertainties of u(T, k=2)= 0.02 K for temperature, u(P, k=2)= 0.0008 MPa for pressure, and u(z) = 0.006 for molar compositions. The VLE data were modeled using the Peng–Robinson equation of state with classical van der Waals mixing rules and an alternative approach combining modified Huron–Vidal mixing rules with the NRTL model for the liquid phase. In addition, the predictive PPR78 and PSRK models were evaluated against the experimental dataset. With optimized binary interaction parameters, all models reproduced the measured data with acceptable deviations, effectively capturing the strongly non-ideal behavior of the H₂S–C₆H₆ system. These results extend the experimental database for H₂S–C₆H₆ mixtures, validate robust EOS-based and predictive modeling frameworks, and provide a reliable foundation for industrial process design, simulation, and optimization
Generative AI impact assessment through a life cycle analysis of multiple data center typologies
No full textInternational audienc
Design specifications for biomedical virtual twins in engineered adoptive cellular immunotherapies
No full textInternational audienceIn (immune)oncology, virtual twins (VTs) offer patient-individual decision support. Nevertheless, current VTs do not incorporate the unique properties of engineered adoptive cellular immunotherapies (eACIs). Here, we outline the minimal design specifications for VTs for engineered ACIs (eACI-VTs) to model the complex interplay between cell product and patient physiology. We motivate utilizing VTs in eACIs to provide decision support and reflect on how eACI-VTs can support the widespread use of eACIs
L'impact de l'Intelligence Artificielle Générative sur les métiers de l'administration : effet gadget ou transformation profonde ?
No full textLorsqu’elle est bien intégrée, pensée avec et pour les métiers, l’IAG peut représenter un formidable levier de transformation du service public. Mais ces apports ne sont pas garantis. Ils dépendent de l’intelligence collective mise dans la conception, le déploiement et l’usage des outils. C’est précisément pour cela qu’une étude fine, contextualisée et critique des usages de l’IAG dans l’administration est nécessaire. Elle permet de lever les malentendus, d’identifier les conditions de réussite, de faire remonter les expériences locales. En somme, elle est une condition indispensable pour « déverrouiller » le potentiel de l’IAG, et éviter qu’elle ne reste un outil sous-utilisé ou mal employé
Effect of dissolved oxygen and temperature on oxidation and stress corrosion cracking of 316L stainless steel in pressurized water reactors primary environment
No full textInternational audienceThe operating experience of 316L stainless steel in pressurized water reactor (PWR) primary environment has revealed a number of cases of intergranular stress corrosion cracking (IGSCC). The primary water is normally deaerated and hydrogenated, but dissolved oxygen can be introduced during shutdowns or water additions, and is postulated to enhance the susceptibility to stress corrosion cracking. Moreover, the primary and auxiliary circuits exhibit a large range of temperatures. The existing literature presents conflicting laboratory findings regarding the impact of both dissolved oxygen and temperature on the susceptibility to stress corrosion cracking (SCC). The aim of this study is to investigate the individual and combined impact of these two parameters on the initiation of SCC in stainless steels exposed to PWR primary water. To this end, different mechanical stress states (slow strain rate tensile tests, constant loading tests, cold worked samples) are employed, and tests are conducted in both hydrogenated and oxygenated environments at different temperatures (290, 320, 340°C). Surface oxides, intergranular oxides (considered a precursor of IGSCC) and SCC cracks depths and lengths are measured to evaluate both intergranular oxidation and SCC susceptibility