HAL-Artois
Not a member yet
27242 research outputs found
Sort by
EnergyCompress: A General Case Base Learning Strategy
No full textInternational audienceCase-based prediction (CBP) methods do not learn a model of the target decision function but instead perform an inference process that depends on two similarity measures and a reference case base. This paper proposes a strategy, called EnergyCompress, to learn an effective case base by selecting relevant cases from an initial set. Use of EnergyCompress decreases CBP inference time, through case base compression, and also increases prediction performance, for a wide variety of CBP algorithms. Ener-gyCompress relies on a general formulation of the CBP task in the framework of energy-based models, which leads to a new and valuable characterization of the notion of competence in case-based reasoning, in particular at the source case level. Extensive experimental results on 18 benchmarks comparing EnergyCompress to 5 reference algorithms for case base maintenance support the benefit of the proposed strategy.</div
« David et Goliath au service du théâtre de marionnettes pour la jeunesse au XIXe siècle »
No full textInternational audienc
« Vraies et fausses infirmités dans le théâtre pour enfants du XVIIIe siècle : une voie vers l’intégration sociale »,
No full textInternational audienc
Sélection génomique du Miscanthus sinensis: Prendre en compte l'effet de l'âge dans la modélisation
No full textInternational audienceMiscanthus, a perennial rhizomatous plant, is of growing interest in France for biomass production and its valorization in the bioeconomy and ecosystem services. However, French cultivation is based on a single variety, which is risky in case of any environmental pressure, and therefore requires the development of new varieties. Recent work of the team has highlighted an increase in heritability over time for most traits of interest (Raverdy et al., 2022), which contributes to increase the selection cycle duration. In this context, the main objective is to develop genomic selection in Miscanthus sinensis in order to shorten its selection cycle. Genomic selection makes it possible to predict the performance of non-phenotyped individuals based on knowledge of their genotypes for numerous molecular markers. Applied to miscanthus, this would make it possible to predict adult performance on the basis of molecular marker genotypes determined at a young age, thus shortening the selection cycle
Impacts environnementaux de différents scénarios de production d'électricité en France
No full textInternational audienceThis study focuses on the environmental impacts of the electricity production scenarios proposed for France for 2060 by the French electricity network manager, RTE. They consider an increasing electricity consumption and a production without fossil fuels, essentially based on wind, photovoltaic, hydraulic, and, for some, nuclear power. The total power to install is significantly higher than the current one; it is higher in the scenarios that rely more on photovoltaic and less on nuclear. Renewable energies require more materials than nuclear, especially photovoltaics. The Simapro Life Cycle Assessment software is used, associated with the Ecoinvent database and CML criteria. Generally, photovoltaic electricity production, which needs more installed capacity, has a more significant impact on the environment, while hydraulic power has less effect. Based on the energy produced as well as on the installed capacity, neglecting the risk of nuclear accidents and nuclear waste management, fossil fuels being excluded from the comparisons, the scenarios that require nuclear power have less environmental impact; those that rely more on photovoltaic power have more environmental impact. For example, CO2 emissions would range from 6.8 to 10.4 g per kWh from one scenario to another. This study can be helpful for policy-makers who have to choose future ways of generating electricity.L'objectif de cette étude est de s'intéresser aux impacts environnementaux des scénarios de production d'électricité proposés pour la France à l'horizon 2060 par le gestionnaire du réseau électrique français RTE. Ils envisagent une consommation d'électricité croissante et une production sans combustibles fossiles, essentiellement basée sur l'éolien, le photovoltaïque, l'hydraulique et, pour certains, le nucléaire. La puissance totale à installer est nettement supérieure à la puissance actuelle ; elle est plus élevée dans les scénarios qui font davantage appel au photovoltaïque et moins au nucléaire. Les énergies renouvelables nécessitent plus de matériaux que le nucléaire, en particulier le photovoltaïque. Le logiciel d'analyse du cycle de vie Simapro est utilisé, associé à la base de données Ecoinvent et aux critères CML. D'une manière générale, la production d'électricité photovoltaïque, qui nécessite une plus grande capacité installée, a un impact plus important sur l'environnement, tandis que l'énergie hydraulique a moins d'impact. Sur la base de l'énergie produite comme de la capacité installée, en négligeant le risque d'accidents nucléaires et la gestion des déchets nucléaires, les combustibles fossiles étant exclus des comparaisons, les scénarios qui font appel à l'énergie nucléaire ont moins d'incidences sur l'environnement ; ceux qui reposent davantage sur l'énergie photovoltaïque ont plus d'incidences sur l'environnement. Par exemple, les émissions de CO2 varieraient de 6,8 à 10,4 g par kWh d'un scénario à l'autre. Cette étude peut être utile aux décideurs politiques qui doivent choisir les futurs modes de production d'électricité
Asymmetrical flow field-flow fractionation and multi-angle laser light scattering: A new analytical approach for the characterisation of insect protein aggregation/polymerisation after heat treatment of Tenebrio molitor larvae
No full textInternational audienceUnderstanding the structural modifications of insect proteins during the transformation processes used for extract preparation is essential for optimising their functionalities and obtaining high added-value proteins. From this perspective and in addition to classical analytical approaches, we developed an original methodology based on the implementation of Asymmetrical Flow Field-Flow Fractionation and Multi-Angle Laser Light Scattering (AF4-MALLS) coupling to quantify and characterise the aggregation / polymerisation phenomena of Tenebrio molitor larvae proteins after heat treatment (from 65 to 95°C). Applied to heat-treated larvae proteins in conjunction with the evaluation of intrinsic fluorescence, surface hydrophobicity and sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE), the AF4-MALLS method enabled us to quantify and characterise the aggregated proteins (forms dissociable after urea denaturation), determining the polymer/monomer (P/M) ratio. Heat treatment significantly affects solubility (-35%), which is due to the amplification of aggregation phenomena, as demonstrated by the increase in the P/M ratio (×10). Moreover, the method enabled us to quantify and characterise the polymerised protein (forms dissociable after chemical reduction of intermolecular bonds), identifying the elements by molar mass and size distribution and conformation.Proteins with cysteine groups can be polymerised under heat, causing a thiol-disulphide exchange reaction and forming a strong (Mw >10 7 g/mol, RGw >130 nm) and compact polymer structure (v ≤0.35) and resulting in intermolecular S-S bonds that preferentially mobilise proteins with Mw >80kDa. Given its performances, the AF4-MALLS method is a real opportunity to understand the effects of processing methods, such as thermal and non-thermal treatments, to optimise protein functionalities.</div
Hydrogen production by photoreforming of carbohydrates catalyzed using gold-based materials
No full textInternational audienc
CO<sub>2</sub> purification of industrial flue gas: effect of physico-chemical properties obtained by different preparation methods on CoCuAl oxide for the Selective Catalytic Reduction process
No full textInternational audienc
An international literature-based dataset on metallic trace element contamination in kitchen garden plants
No full textInternational audienceUrban agriculture is an attractive solution to counter artificialization and densification of (sub)urban areas, offering environmental, socio-economic and health benefits. Nevertheless, the transfer of metallic trace elements (MTE) from soil, water or air to crops raises concerns about potential health risks for consumers in urban environments. Assessing environmental suitability is therefore essential for safe food production. To support this, the BAPPET dataset compiles data on MTE concentrations in 90 edible plants commonly grown in kitchen gardens. Built from 528 international studies, the dataset also describes MTE content in the environment and key parameters influencing MTE transfer to plants: (i) Plant information: species, variety, cultivation conditions, harvesting details; (ii) Growing media: MTE concentrations in soil, air and water, agronomic soil parameters affecting MTE availability; (iii) Experimental context: MTE contamination sources, methodological details. With over 29,000 analyses, the BAPPET dataset provides a comprehensive data collection for the comparison of site-specific results in environmental diagnoses and for the construction and validation of transfer models