HAL Portal AU (University of Avignon)
Not a member yet
35713 research outputs found
Sort by
Learning Market Equilibria Preserving Statistical Privacy Using Performative Prediction
No full textInternational audienceprivate network game, where end users minimizetheir cost by computing their demand and controllable generation.Their nominal demand constitutes sensitive informationthat they might want to keep private. We prove that theprivate network game admits a unique variational equilibrium,which depends on the private information of all end users.Thus, to update their strategy, end users rely on randomizedreadings. A data aggregator is introduced, which aims to learnthe end users’ private information, while remunerating themdepending on the quality of their readings. Using performativeprediction, we define a decision-dependent game explicitlytaking into account the distribution shift caused by the endusers’ hidden ability. The decision-dependent game coincideswith a Stackelberg game when the end users’ hidden abilitiesare best responses. Further, the market robustness can bequantified by evaluating the efficiency loss as the differencebetween the social cost in the performatively stable equilibriumand the optimum. We show that under mild assumptions, theperformatively stable equilibrium can be found by distributedand sequential variants of the repeated stochastic gradientmethod while we propose a two-timescale stochastic approximationmethod to learn Stackelberg equilibrium. Finally, weformulate the data aggregator’s optimal contract design as abilevel optimization problem that we cast as a more tractablenonlinear nonconvex optimization problem which can be solvedusing simulated annealing. Simulations on small and large scaleproblem instances illustrate the results
Let's get functional: Relationship between earthworm traits and physicochemical cast properties
No full textInternational audienceAlthough earthworms play a crucial role in soil biogeochemical processes, the importance of their traits in shaping the physicochemical properties of their casts remains poorly understood. This study aimed (1) to evaluate the influence of earthworm species and soil types on cast properties and (2) investigate the relationship between earthworm morphological, anatomical, physiological and behavioral traits and the physicochemical properties of their casts. Nine temperate earthworm species were introduced under controlled conditions in two contrasting soil types (Luvisol and Cambisol), and 21 traits were determined for each species. Casts were analyzed for their physicochemical properties, total and available organic carbon and nitrogen, and compared to control soil incubated under similar conditions without earthworms. Results showed that earthworm activity changed soil pH depending on soil type, with pH increasing in Cambisol, and decreasing in Luvisol. Species-specific effects on cast properties revealed a physicochemical gradient: the epigeic L. castaneus and epi-anecic L. terrestris produced casts with the highest nitrate, dissolved organic carbon, total organic carbon, and moisture levels, whereas cast from endogeic species showed the lowest values. Moreover, earthworm species exerted a stronger overall influence on cast properties than soil type (59 % vs. 24 %), underscoring the dominant role of species-specific traits in shaping cast characteristics. We identified nine key traits related to the earthworms' morphology, anatomy, physiology and behavior, that influenced cast properties directly or indirectly. Direct effect traits included mouth area, gizzard size and location, typhlosole complexity, intestinal mucus and cast production. Indirect effect traits, such as pigmentation, litter ingestion and litter-to-cast ratio, reflected the ecological behavior of earthworm species. This trait-based approach provides a promising avenue for future studies on the role of earthworms in soil biogeochemical cycling and a framework for improving our understanding of their impacts on soil organic matter dynamics
Insights into the mechanisms involved in the evolution of the structural and physicochemical properties of tomato during air drying – A study combining MRI, unilateral NMR and conventional techniques
No full textInternational audienceMost fruits and vegetables have high water content, making them highly perishable. Drying preserves them by reducing water content, slowing microbial activity and chemical spoilage. It also enhances product range and reduces transport and storage costs. This study explored how water and air distribution and chemical properties of fruits affect the drying process and quality of dried tomatoes, focusing on texture and nutrient composition. MRI and unilateral NMR, the unique spatially resolved techniques to characterize water dynamics and multiscale tissue structure non-destructively, were combined with conventional techniques assessing chemical composition, histological features and texture properties.MRI and unilateral NMR succeeded in capturing tissue evolution and heterogeneous features of tomato slices during drying by accessing information on water status and distribution and apparent micro-porosity. Results indicated that fruit microstructure influenced drying kinetics: larger cells and less heterogeneous tissue promoted dehydration. Drying also changed biochemical compounds and phenolic profiles, affecting nutrient quality. Despite differences in drying rates between the two cultivars investigated, their texture and nutrient composition remained similar, suggesting potential for improvement of cultivars destined for drying and/or of the drying process itself. These findings could help optimize drying processes and develop tomato cultivars better suited to drying. Last, the experimental approaches adopted here, i.e., MRI for in situ drying investigation and unilateral NMR, can help gain deeper insights into fleshy fruit transformation, with possible industrial applications.</p
Syntactic Study of Self-Repair and Self-Reformulation in French and Spanish: Effects of Utterance Length
No full textIn this article we propose a study of self-repair and self-reformulation based on the analysis of a corpus of nearly 19,000 tokens, fully segmented into utterances. This corpus includes two languages, French and Spanish, which are equally represented, and five different communicative contexts (sociolinguistic interview, public interview, work meeting, meeting between friends and service interaction) for each of these two languages. This corpus architecture, as well as the segmentation of the entire corpus into utterances, allows us to point out some general trends, not previously described, regarding the effect of unit length on self-repair. Indeed, the average length in tokens of the utterance not only predicts the proportion of units with at least one self-repair, but also models the site of initiation of self-repair and type: if less than 10% of utterances with 5 tokens or less have self-repair, almost 40% of utterances with between 11 and 20 tokens have self-repair, and about 70% of utterances with more than 31 tokens have self-repair. Moreover, in utterances with 5 tokens or less, self-repair is most likely to be without reformulation and to occur in a unit without a predicate, whereas in utterances with more than 11 tokens self-repair is likely to be with or without reformulation and to be initiated after the predicate, towards the middle of the unit
Sanguis qui est Mantue : les traités mantouans De sanguine Christi de Battista Spagnoli et Pietro Gavassetti (1492)
No full textInternational audienc
Enjeux épistémologiques et éthiques d’une recherche-action participative dans le développement professionnel d’enseignants d’anglais en CAP
No full textInternational audienc
Modelling the Road Network as an Expression of Historical Spatial System Change in the Klaipeda Region (Lithuania)
No full textInternational audienceSince the 20 th century, the Klaipėda region (Lithuania) has undergone significant political, economic, social, and cultural changes related to the disappearance of Prussia (1923), the Soviet occupation (1944) and the restoration of independence (1991). This has led to radical transformations in the spatial system, of which the road network is a key component. Networks could be analysed using space syntax methods by assimilating them to a mathematical graph model, for which quantitative indexes based on topology can be applied to assess their spatial configurations. The 'Generic City' approach, by simulating foreground and background urban networks, can be modified innovatively on a territorial scale. Based on historical cartographies and open geographic databases, indexes characterising both networks are derived to describe the length, angularity or accessibility of the pre-Soviet, Soviet and post-Soviet networks. The analysis identified four main spatial structures associated with the network by combining the indexes in a K-means machine learning algorithm. They highlight the spatial impacts of collectivisation, industrialisation, and tertiarisation of the economy, post-World Wars and post-Cold War geopolitical events, and their consequences as drivers of the territorial organisation and dynamics such as the metropolisation or peri-urbanisation
LED UV-A/chlorine filterless domestic purification system for the elimination of emerging organic contaminants and faecal bacteria indicators in tap water
No full textInternational audienceThe development of new purification systems is a crucial issue to remedy water contamination concerns particularly in the worldwide context of water quality deterioration and scarcity. Therefore, a newly commercialized domestic water purification device, using LED UV-A/chlorine advanced oxidation process, was tested for the first time in representative matrix conditions to improve drinking water quality. The removal of 11 emerging organic contaminants (pesticide, hormone, endocrine disruptor, pharmaceutical, perfluoroalkyl substances (PFAS)) and of 4 faecal indicators strains (Escherichia coli and Enterococcus faecalis) was explored through a general degradation kinetical study. The experiments were performed in different conditions (dark, UV-A, [chlorine] = 0, 0.2, 1 and 3 mg L−1) and elimination and disinfection rates were determined as well as revivification potentials. LED-UVA/chlorine degradation was observed in the range 2 to 100 % for all emerging organic contaminants depending on chlorine concentration except for PFAS. The degradation is mainly attributed to oxidation by hydroxyl radicals for most of the compounds and to chlorination for bisphenol A and 17-α-ethynyl-estradiol. In all tested conditions, a total inactivation of Escherichia coli strains was observed, whereas Enterococcus faecalis strains showed more resistance under LED-UVA irradiation. LED-UVA irradiation proved to alter cellular functions preventing revivification after treatment
« Reading landscape » 讀景 in a Mediterranean peri-alpine catchment basin: (Vallée des Duyes, Geopark UNESCO of Haute Provence, Alpes de Haute Provence, France)
No full textInternational audienceNota bene: this oral presentation in English is based on an article written in French that will be published in September 2025 Christian Depraetere, Benoît Deffontaines, Jiun-Chuan Lin, Jean-Simon Pagès, 2025. Évènements hydroclimatiques passés et actuels dans le bassin versant de la vallée des Duyes (Alpes de Haute-Provence). Revue Méditerranée, accepté pour publication.Located in the UNESCO Geopark of Haute Provence, the Duyes watershed is representative of the hydrodynamic processes of the Mediterranean peri-alpine mid-mountains with deformed sedimentary bedrock. The local archives record a major hydroclimatic event associated with ten days of heavy rainfalls affecting this valley in April 1774 with a large-scale landslide with several damages: sinking of arable land, catastrophic flooding, destruction of buildings and roads. The archives describe the events in the terms of the time: "continuous rainfall", "abundance of water from the torrents", "revolution of the earth", "openings and crevasses", etc. The documents also point out that these "strange inconceivable revolutions" were concomitantly associated with two earthquake tremors without stating if this was a cause or a consequence of the detachment of several millions tons of rock and mud masses. This event took place at the end of the Little Ice Age (LIA, 1350-1850) in the context of a densely populated rural world practicing traditional peasant polyculture in the form of mid-mountain extensive agropastoralism.The first part presents and comments the 1774 archives as an example of a geohistorian's work. The second part details the climatic and hydrogeomorphological contexts of the valley at the time based on field observations and current data, including morphometric mapping using the LIDAR digital terrain model (DTM) and torrential flood simulations. Extensive fieldworks and the expertise of the local population also form the backbone of this study. Current flood and landslide risks are briefly recalled, based on the climatology of the years 2022 to 2024, in particular in May 2023 when heavy rainfall caused gullying of dirt roads and landings at the foot of steep fields and pastures in many valley locations.This research aims to contribute to the geohistory of the environmental and human societies co-evolution, and more specifically on the interactions between hydrogeomorphological processes and the dynamics of anthropized agrofestry landscapes in Mediterranean mountain regions. It also stresses the value of practical training in “landscape reading”, following the pedagogical principle of 讀景 put forward by Professor Jiun-Chuan Lin in his magnum opus on applied geomorphology (地形學,2022)
