Portail HAL Institut Agro Montpellier
Not a member yet
54878 research outputs found
Sort by
The impact of european AI regulation: Governance and strategic responses in Switzerland and South Korea
No full textInternational audienceThis research project explores how the EU AI Act and its related measures are influencing the development of AI governance, regulation, and strategy in non-European countries, with a focus on Switzerland and South Korea. Both nations are at the forefront of IT innovation. Switzerland, despite not being an EU member, maintains close ties with the EU through bilateral agreements and partial access to the single market, while South Korea, though geographically and politically distant, engages with the EU through strategic partnerships and trade relations.The project seeks to understand how the EU AI Act impacts national AI discourses in these two countries. It will analyze the discourses and imaginaries being shaped around AI governance, identify the key proponents of these discourses, such as government agencies, advisory bodies, and private industry leaders, and compare the strategies employed to adapt to or diverge from the EU's regulatory framework. In terms of temporality, it distinguishes three periods: before 2019, 2019-2023 (marked by the Coronavirus outbreak and the subsequent economic recovery plans), and after 2024 (characterized by the emergence of AI services based on Large Language Models and the endorsement of the European AI Act).The methodology involves qualitative documentary analysis of policy documents, white papers, reports published by government bodies and advisory institutions, and media coverage of these elements. This will be conducted both manually and through computational textual analysis techniques such as text mining and topic modeling. These methods will help identify recurring themes, dominant narratives, and strategic differences between Switzerland and South Korea in their adaptation to or divergence from EU AI regulations, offering insights into how the EU’s regulatory influence affects AI governance in regions both near and far
The construction of an interdisciplinary field: The case of digital agriculture in France
No full textInternational audienceThis paper investigates whether a strategically organized, regionally focused investment in higher education can accelerate both the promotion of an innovative socio-technological field and the stabilization of its research community. It focuses on a Convergence Institute for Digital Agriculture (DA Institute), launched in 2017 to advance digital agriculture in Montpellier, France. Known for its interdisciplinary approach, the DA Institute united multiple universities and research institutions and funded around 70 doctoral theses, all requiring interdisciplinary collaborations. These efforts aim to foster innovation in digital agriculture, a field that integrates technology with traditional farming practices to address current societal challenges.The study aims to assess whether the DA Institute has successfully shaped digital agriculture into a distinct research field centered in the region and fostered collaborative networks that extend beyond formal co-supervision relationships. Additionally, it explores if this initiative has contributed to establishing a unique identity for digital agriculture, distinct from similar efforts in other countries. Furthermore, the research examines the broader impact of the DA Institute on the scientific landscape and interdisciplinary collaborations in the region, comparing research funded by the DA Institute with independent research projects. The central question is whether this strategic investment has accelerated the development of the field and contributed to a stable research community.To explore these questions, this paper draws on a qualitative documentary analysis of policy documents related to the DA Institute and interviews with participating researchers and PhD students. Additionally, a bibliometric study of publications produced by these researchers will track the evolution of keywords and scientific collaborations. This combination of qualitative and bibliometric approaches will provide a comprehensive evaluation of the effects of the DA Institute's strategic investment on the development of digital agriculture and the construction of a cohesive research community
N RAMP2-Mediated Manganese distribution in plant tissues: A New Role in the Seed
No full textInternational audienc
Postface. Apprendre la forêt en la dégustant
No full textSource Agritrop Cirad (https://agritrop.cirad.fr/615977/)International audienc
Machine-Actionable Metadata in Practice: Lessons From Automating FAIR Assessment in Plant-Pollinator Datasets
No full textMISTEA - Axe informatiqueInternational audiencePlant-pollinator interactions play a pivotal role in ecosystem functioning and sustainable agriculture. However, plant-polinator datasets are scattered across various networks, in country-specific initiatives, and stored in isolated silos, making them difficult to access by scientists and decision-makers. By promoting the adoption of Findable, Acessible, Interoperable, and Reusable (FAIR) data standards (Wilkinson et al. 2016) across multiple initiatives worldwide, we are working to transform the fragmented nature of these datasets and make data on plant-pollinator interactions widely available. As the biodiversity community advances towards FAIR data, machine-actionable metadata has emerged as a critical enabler for scalable data assessment, discovery, and reuse. However, while FAIR principles emphasize machine-readability, many datasets are still evaluated manually or lack structured metadata entirely, limiting their integration into global platforms. This study shares practical insights from the WorldFAIR Agricultural Biodiversity Case Study, in which we operationalized machine-actionable FAIR metadata for the review of plant-pollinator interaction datasets (Drucker et al. 2024). We developed a semi-automated workflow to assist in evaluating datasets against the FAIR principles using tools from the Global Biotic Interactions initiative (GloBI, Poelen et al. 2014). The GloBI bots "Nomer" and "Elton" can read structured metadata from standard vocabularies such as Darwin Core (DwC), Ecological Metadata Language (EML), and the Plant-Pollinator Interactions (PPI) vocabulary. Nomer focuses on taxonomic alignment with several taxonomic catalogues, such as GBIF Backbone and Catalogue of Life. Elton extracts species interactions from datasets of various structures and formats, including DwC-Archives. By relying on machine-readable metadata, the bots were able to flag inconsistencies, suggest improvements, and generate repeatable reports across pilot projects in Argentina, Brazil, the African continent, Kenya, Colombia, East Africa, Central Asia, and the USA (for example, Elton et al. 2025). This helped researchers assess dataset interoperability without needing full access to the raw data, a crucial feature given legal and institutional access constraints. To make the data review report readable for researchers, GloBI's bots produce a document resembling a data publication, complete with a title, authors, publication date, abstract, introduction, and other relevant sections. Our results underscore the transformative role of machine-actionable metadata in biodiversity data governance. Automating FAIR assessments not only increases transparency and repeatability but also accelerates the integration of datasets into platforms like GloBI. While human expertise remains essential, tools like Nomer and Elton demonstrate that FAIR assessment can evolve beyond checklists to become dynamic, scalable, and integrated into the data lifecycle. Resources and code are openly available in the online repositories Zenodo*3 and GitHub*2, and are summarized in the GloBI platform*1. To help alleviate the burden of manually reviewing data as part of scientific publication review, we propose deploying domain-specific, automated data review processes that enable researchers to better understand how to make their data easier to review and reuse. Recognizing that publishing reusable, integrated data remains mostly a manual process, we recommend that plant-pollinator and species interaction datasets be registered with one or more infrastructures (e.g., GloBI, GBIF) to benefit from the domain-specific data review services they offer. We suggest that data publishers continue to collaborate on building, maintaining, and improving similar infrastructures to assess and increase the quality and FAIRness of published scientific data. Through the adoption of standards such as Ecological Metadata Language, Darwin Core, Plant-Pollinator Interactions Vocabulary, and Relation Ontology, we aim to enhance the understanding of how plant-pollinator interactions contribute to sustaining life on Earth while ensuring that data is easily findable, accessible, interoperable, and reusable for further research and analysis (FAIR)
An empirical analysis of the determinants of food insecurity among smallholder farmers in Eastern Rwanda
No full textSource Agritrop Cirad (https://agritrop.cirad.fr/615956/) * Autres projets (id;sigle;titre): DCI-ENV/2017/387-627;;(EU) Regreening Africa//International audienceBackground: Food insecurity is one of the most pressing problems confronting households in sub-Saharan Africa (SSA). It is particularly acute in low income areas across the continent. Despite Rwanda's economic progress, food insecurity persists especially in rural areas, necessitating empirical evidence to inform targeted interventions that address the complex interplay of socio-economic, environmental and policy factors affecting household food insecurity. Factors affecting food insecurity vary, and obtaining context-specific information is necessary for designing relevant interventions. This empirical study analyzes factors affecting the probability of experiencing severe food insecurity in Eastern Rwanda and discusses how land restoration strategies like agroforestry may contribute to it. Panel data collected in 2018 and 2022 from 1100 randomly selected households are analyzed using both descriptive statistics and a correlated random effects probit model. Results: The findings show a generally high level of food insecurity, with sample households having average food insecurity experience scale scores of 6.02 in 2018 and 5.73 in 2022. Moreover, 63% and 60% of the households experienced severe food insecurity in the two periods, respectively. The empirical results show that farming practices and household socio-economic characteristics played a more significant role in food insecurity status. Households that cultivated different crops had a lower probability of experiencing severe food insecurity and larger households were more likely to experience severe food insecurity. However, agroforestry-related variables were not statistically significant in reducing the probability of severe food insecurity experience in the study area. Conclusion: The study concludes that food insecurity in the study area is high. To address the prevailing situation, efforts to reduce food insecurity should focus on solutions that could increase food production in the short term, such as improving household socio-economic status, diversifying crop production and market-focused production. However, these need to be aligned with local needs and ecological conditions. Agroforestry interventions should focus on integrating suitable tree species into farming systems, and future studies should account for the time dimension to accurately capture long-term effects of such interventions. Moreover, experimental studies to enable rigorous impact analysis of agroforestry interventions are recommended
Recherches sur le pastoralisme en France : état des lieux des connaissances et questions vives
No full textInternational audienceThis synthesis provides a state of play of pastoral systems and territories in France, through the lens of the main challenges they face. Particular attention is given to pastoral decline under conditions of production intensification and herd expansion, adaptation to climate change and market fluctuations, and the return of wild predators. Five key areas of research have been defined: animal selection and breeding in pastoral environments; pastoralism as a specific agroecological model, with its strengths and weaknesses; multi-stakeholder pastoral territories, as spaces of confrontation and of collective project development; pastoral professions and their attractiveness; and data derived from monitoring methods for tracking changes in vegetation, biodiversity, and livestock farming systems. Many of these issues show similarities with pastoral situations observed in other regions of the world, notably in West Africa, without entirely overlapping with them.Cette synthèse dresse un état des lieux des connaissances sur les systèmes et territoires pastoraux en France au prisme des principaux enjeux qui les traversent. Le repli pastoral en situation d’intensification de la production et d’agrandissement des troupeaux, l’adaptation au changement climatique et aux fluctuations des marchés, ou encore le retour de prédateurs sauvages sont particulièrement discutés. Cinq champs de recherche prioritaires ont été dégagés : l’animal en milieu pastoral ; le pastoralisme comme modèle d’agroécologie avec ses forces et ses faiblesses ; les territoires pastoraux multi-acteurs, espaces de confrontation et de projets collectifs ; les métiers du pastoralisme et leur attractivité ; les données issues des méthodes de suivi de l’évolution des végétations, de la biodiversité et des systèmes d’élevage. Bien des sujets font écho avec les situations pastorales du reste du monde, notamment en Afrique de l’Ouest, sans pour autant les recouvrir exactement
Coexistence et confrontation entre exploitations bio et non-bio dans les Cévennes Gardoises: Focus sur la haute vallée de l'Hérault et la haute vallée Borgne
No full textThis report is the result of a research at INRAE Montpellier's Innovation Research Unit. It focuses on the phenomena of coexistence and confrontation between organic and non-organic farms in the Cévennes region of Gard, more specifically in the Causses Aigoual Cévennes community of municipalities. The internship is part of the DEFIBIO project, which aims to address the challenges facing organic farming in Occitanie. In this context, the question is not only whether organic and non-organic farming can coexist, but rather how this coexistence is built in concrete terms, around what objects (resources, practices, representations) and with what effects on agricultural and territorial trajectories. Through 27 qualitative interviews with sheep and goat farmers, arborists, diversified market gardeners, and sweet onion producers, we analyzed the interactions between farmers and the factors that structure them. We show that organic farming is not only a criterion for distinguishing between producers, particularly in the case of onions, but also a factor that structures professional and social relationships, which combines with other divisions (sectors, practices, social trajectories, ideologies) to redraw the lines of coexistence in the Cévennes region of the Gard department. Coexistence therefore involves social, economic, and political relationships, in which organic farming plays a structuring but never exclusive role.Ce rapport est le fruit d'un stage de recherche au sein de l'UMR Innovation d'INRAE Montpellier. Il s'intéresse aux phénomènes de coexistence et de confrontation entre exploitations agricoles biologiques et non-biologiques dans les Cévennes gardoises, plus précisément sur le territoire de la communauté de communes Causses Aigoual Cévennes. Le stage s'inscrit dans le projet DEFIBIO, qui a pour objectif de relever les défis auxquels l'agriculture biologique fait face en Occitanie. Dans ce contexte, la question n'est pas seulement de savoir si les agricultures bio et non-bio peuvent coexister, mais plutôt comment cette coexistence se construit concrètement, autour de quels objets (ressources, pratiques, représentations) et avec quels effets sur les trajectoires agricoles et territoriales. Au travers de vingt-sept entretiens qualitatifs avec des éleveurs ovins et caprins, des arboriculteurs, des maraîchers diversifiés et des producteurs d'oignons doux, nous avons analysé les interactions entre agriculteurs et les facteurs qui les structurent. Nous montrons que l'agriculture biologique ne constitue pas seulement un critère de distinction entre producteurs, notamment dans le cas de l'oignon, mais bien un facteur structurant des relations professionnelles et sociales, qui se combine à d'autres clivages (filières, pratiques, trajectoires sociales, idéologies) pour redessiner les lignes de coexistence dans les Cévennes gardoises. La coexistence engage donc des rapports sociaux, économiques et politiques, où le bio joue un rôle structurant mais jamais exclusif
What sort of digitalization do family farmers need ?
No full textSource Agritrop Cirad (https://agritrop.cirad.fr/615899/) * Autres projets (id;sigle;titre): ;;(FRA) Projet Fracture Numérique// ;ACOTAF;(FRA) Renforcer le conseil agricole pour accompagner les transitions agroécologiques de l’agriculture familiale en Afrique subsaharienne// ;Compairs;(FRA) Certification par les pairs pour une qualité éco-solidaire : co-construction d’un commun intellectuel// ;DigiCLA;(BEN) Le digital pour lutter contre la Chenille Légionnaire d'Automne//International audienceDigitalizing the agricultural sector in West Africa is a promising opportunity for stakeholders in the agricultural and technology sectors, an opportunity that relies on agricultural producers' voluntary mass appropriation of connected phone tools. Mobile phones and app use could foster independent producer networks, improve value chain structuring and boost the sector's economic value. Agricultural producers are now being targeted by a growing number of providers of “digital services”, in the form of apps developed by tech and agritech start-ups. However, apps developed specifically for agricultural producers have met with mixed success. What lessons can be drawn from this mixed bag of results, and what will it take for digitalization to truly serve producers and facilitate the transition to socially, economically and environmentally sustainable practices
Adapting BERT and AgriBERT for Agroecology: A Small-Corpus Pretraining Approach
No full textInternational audienceSource variables, or observable properties, used to describe agroecological experiments are often heterogeneous, non-standardized, and multilingual, making them challenging to understand, explain, and utilize in cropping system modeling and multicriteria evaluations of agroecological system performance. A potential solution is data annotation via a controlled vocabulary, known as candidate variables, from the Agroecological Global Information System (AEGIS). However, matching source and candidate variables via their textual descriptions remains a challenging task in agroecology. Domain-general language models, such as BERT, often struggle with domain-specific tasks due to their general-purpose training data. In the literature, these models are adapted to specialized domains through further pretraining, pretraining from scratch, and/or fine-tuning on downstream tasks. However, pretraining a domain-general model on a domain-specific corpus is resource-intensive, requiring substantial time, energy, and computational resources. To the best of our knowledge, no study has further pretrained a domain-general model on a small corpus (less than 100 MB) to adapt it to a domain-specific task and evaluated it on downstream tasks without fine-tuning. To address these shortcomings, this paper proposes further pretraining BERT and AgriBERT on a small agroecology-related corpus. This approach is designed to be both time- and resource-efficient while enhancing domain adaptation. We evaluate the pretrained models on the task of matching source and candidate variable descriptions without fine-tuning. Our results show that our further pretrained AgriBERT (+ Experts + Core) model outperforms all others by more than 8% from P@1 to P@10. These findings showed that small-scale pretraining can significantly improve performance on domain-specific tasks without requiring fine-tuning