Examinando por Autor "Lezoche, Mario"
Mostrando 1 - 10 de 10
Resultados por página
Opciones de ordenación
- Resumen
Acceso Abierto Agro-Knowledge Integration: Developing a FAIR data science approach for adding value to the agricultural supply chain(2021) Torres, Diego; Lezoche, Mario; Collazos, César; Codocedo, Víctor; Motz, Regina; Antonelli, Leandro; Panetto, Herve; Fernández, AlejandroFarms are the engine to support rural employment making a considerable contribution to territorial development. Even though they have always been considered a cornerstone of agricultural activity in the European Union (EU) and in Latin America, this sector most often suffers from very low efficiency and effectiveness, sensitivity to weather, market disruptions and other external factors. Two different problems in knowledge sharing are present in this domain. First, the various interoperability regulations between the countries. Although some efforts are done to bypass this problem, like the EU-Mercosur signed in the summer of 2019, the different process semantics implemented in each region are a serious threat to the fulfillment of the process interoperability. Another problem is that in most of the cases, the knowledge transferred from generation to generation is paramount from a cultural point of view, but most of the time, it does not answer to the needs nor the requirements of the agri-food value chain. We aim at creating the core technology for a knowledge hub that integrates and aligns international regulations in agricultural activities, such as FAO's best practices, and possibly the last-born EU-Mercosur regulations with the local restrictions, such as national policies, allowing the small farmers to access, in an easy way, a wider market through the certification of the practices and products. In order to develop this core technology, we propose to deploy various methodologies and tools working on the domains of knowledge formalization, domain alignment and visualization. The domain of formal representation allows for the semantic alignment of rules and restrictions from different institutional regulation bodies. Simultaneously, we will propose a model for incoherence detection letting us to highlight contradictory regulations. Those knowledge atoms and constructs will be represented through some visualization information interfaces according to the users’ needs. The methods and tools that will be employed are at the same time the pillars from the multi relational data mining (MRDM), the artificial intelligence (AI), the knowledge formalization (KF) domains, but will extend the interoperability properties of those domains to become a new interesting and valuable tool for the presented problem. This abstract is issued from an accepted Stic-AmSud project that wad elaborated during the secondments of the RUC-APS project. - Documento de conferencia
Acceso Abierto Extending RCA algorithm to consider ternary relations(2022) Leutwyler, Nicolás; Lezoche, Mario; Panetto, Hervé; Torres, DiegoRelational Concept Analysis (RCA) is a multirelational data mining method that aims to extract knowledge from multiple formal contexts (i.e., objects, attributes, and a binary relation between them) and the relations between them. One of the problems RCA has is the lack of the possibility of extracting knowledge directly from data that is represented with ternary relations. While there are some existing solutions towards this problem, either they require complex preprocessing of the input data, or they lose some capabilities of RCA such as the different meanings of the relations between concepts (∃, ∀, etc). In this work, we present an intuitive extension to RCA to be able to use it with data directly represented with ternary relations. As an example of its usage, we apply it to a dataset called Knomana which includes ternary relations - Artículo
Acceso Abierto Generic software for benchmarking Formal Concept Analysis: Orange3 integration(2023) Leutwyler, Nicolas; Lezoche, Mario; Panetto, Hervé; Torres, DiegoThanks to the internet of things (IoT) and cyber physical systems (CPS), we face an incremental growth of the available data, either on the internet or in private databases. This resulted in data mining techniques becoming an essential piece in the information retrieval process. Moreover, trends like the industry 4.0 encourages its usage to support data driven decisions, for instance. Formal Concept Analysis (FCA) is one of the most used techniques in the unsupervised data mining field due to its inherent ability to find patterns between concepts. As a consequence, many applications need the use of fast algorithms to perform the calculations to retrieve either the lattice or the association rules related with the data at their disposal. Due to this, scientists often rely on manually crafted benchmarks to compare how certain algorithms perform under different circumstances. In this work, we propose the architecture of a software to generalize these benchmarks independently of the algorithms, to be integrated in the open source data analysis software Orange3. - Resumen
Acceso Abierto Knowledge discovering from multiple sources in agriculture value-chain(2021) Torres, Diego; Lezoche, Mario; Fernández, Alejandro; Antonelli, Leandro; Panetto, HerveThe agri-food value-chain results from the interaction of multiple stakeholders. Each stakeholder contributes with a distinct perspective and interest. The diversity in activities and work forms in the value-chain results in a wide variety of data sources, and data management practices. It is common to find information managed in databases, document repositories, or even social media. Document formats also vary (e.g., CSV, PDF, XML, etc.), and so do content types (e.g., graphics, tables, lists, images, etc.). In this context, effective decision-making relies heavily on the availability of interoperable, comprehensive, accurate, and timely information. Knowledge graphs (KG) are graphbased data models for knowledge extraction from multiple structured and unstructured sources that support multilingual integration. KG are frequently combined with knowledge discovering approaches like embedding and multi-relational data mining methods like the Formal Concept Analysis (FCA) and its extension the Relational Concept Analysis (RCA). This work proposes an automatic pipeline process to combine and align different agri-food information sources to discover new pieces of knowledge based on KG and RCA. The approach combines several research lines: (1) entities and relations detection in different sources; (2) alignment with a shared ontology description, based on GACS and AGROVOC, and (3) discovering new knowledge with Relational Concept Analysis in the shape of association rules formalized following the description logic. - Documento de conferencia
Acceso Abierto Knowledge mobilisation crossing boundaries: a multi-perspective framework for agri-food value chains(2020) Liu, Shaofeng; Zhao, Guoqing; Chen, Huilan; Fernández, Alejandro; Torres, Diego; Antonelli, Leandro; Panetto, Hervé; Lezoche, MarioKnowledge has long been recognised as a valuable asset to individuals, organisations and economy, subsequently knowledge management (KM) has been a well-established area of research. Existing research has developed various classification schemes for knowledge, and a great number of KM process and lifecycle models have been proposed over the last few decades. In particular, knowledge transfer and sharing has received great attention. However, majority of existing work has focused on knowledge sharing within the same organisation or community where people have a shared sense of identity, values and some common practice, hence knowledge process and learning is within relatively homogeneous groups. There is inadequate research to address the issue of knowledge boundaries and approaches to knowledge mobilisation spanning across knowledge boundaries. The knowledge boundaries can erect significant barriers to knowledge sharing and flowing especially in value chain context where there are a great number of players from different domains, with varied level of knowledge, having different and possibly conflicting interests - participating in knowledge sharing activities. This paper will explore how such knowledge boundaries can be identified and how knowledge gaps among different value chain players can be closed by using appropriate boundary-crossing mechanisms. A multi-perspective knowledge mobilisation framework is proposed. Example applications of the knowledge mobilisation framework in agri-food value chain will be illustrated, based on the most recent developments from an EU collaborative project, RUC-APS (standing for Risk and Uncertain Conditions in Agriculture Production Systems), which is funded by European Commission’s Horizon 2020 RISE programme. - Documento de conferencia
Acceso Abierto A Method to obtain a Knowledge Representation from a Natural Language Specification of the Domain using the Glossary LEL(2023) Antonelli, Leandro; Lezoche, Mario; Delle Ville, JulianaGood requirements (correct, consistent, unambiguous, etc.) are crucial to software development success. Errors made in the requirements stage can cost up to 200 times if they are discovered once the software is delivered to the client. Natural language artifacts are the most used tool to write requirements, since they are understandable by the both parties that participate in the software development: the stakeholders and the development team. Nevertheless, natural language can introduce many defects (ambiguity, vagueness, generality, etc.). Formal reasoning is a good strategy to check whether requirements satisfy the attributes of good requirements or not, but formal reasoning cannot be applied to natural language specification with defects. Thus, this paper proposes an approach to write a good specification and obtain knowledge from it. The approach uses a particular lexicon, the glossary LEL, and it suggest guidelines to write good specification, and it also suggest rules to obtain knowledge (concepts and relations) from the glossary LEL. The paper also presents a prototype to assist to this approach, and a preliminary evaluation of the approach. - Revisión
Acceso Abierto Methods for concept analysis and multi-relational data mining: a systematic literature review(2024) Leutwyler, Nicolás; Lezoche, Mario; Franciosi, Chiara; Panetto, Hervé; Teste, Laurent; Torres, DiegoThe Internet of Things massive adoption in many industrial areas in addition to the requirement of modern services is posing huge challenges to the field of data mining. Moreover, the semantic interoperability of systems and enterprises requires to operate between many different formats such as ontologies, knowledge graphs, or relational databases, as well as different contexts such as static, dynamic, or real time. Consequently, supporting this semantic interoperability requires a wide range of knowledge discovery methods with different capabilities that answer to the context of distributed architectures (DA). However, to the best of our knowledge there is no general review in recent time about the state of the art of Concept Analysis (CA) and multi-relational data mining (MRDM) methods regarding knowledge discovery in DA considering semantic interoperability. In this work, a systematic literature review on CA and MRDM is conducted, providing a discussion on the characteristics they have according to the papers reviewed, supported by a clusterization technique based on association rules. Moreover, the review allowed the identification of three research gaps toward a more scalable set of methods in the context of DA and heterogeneous sources. - Documento de conferencia
Acceso Abierto Multi-relational and Concept Analysis based Knowledge extraction in the Industry 4.0: A systematic mapping(2023) Leutwyler, Nicolás; Lezoche, Mario; Torres, Diego; Panetto, HervéSmart Enterprises, Smart Manufacturing, and Cyber-Physical Systems are gaining traction in many industry areas. On top of that, the amounts of available data grow rapidly, and organizations are eager to exploit their advantages. To accomplish that, it is mandatory to have a wide variety of methods and algorithms for knowledge extraction in order to fit the different needs and problems of the industry. In this study, we review and dissect the current state of the art in knowledge extraction applied to smart enterprises, smart manufacturing, and cyber-physical systems. More specifically, we provide a classification of the characteristics of the available methods in the literature according to their applications, and point out areas of improvement. - Documento de conferencia
Acceso Abierto Scenarios, shared understanding, and group decision support to foster innovation networks(2023) Agredo-Delgado, Vanessa; Antonelli, Leandro; Collazos, César A.; Fernández, Alejandro; Zaraté, Pascale; Camiller, Guy; Hurtado, Julio; Lezoche, Mario; Motz, Regina; Panetto, Herve; Torres, DiegoCollaborative innovation involves diverse individuals and organizations working together to develop new ideas, products, or services. Successful collaboration in networked innovation projects is challenging due to the need to cross the knowledge boundaries that exist between organizations, disciplines, and cognitive frames. We propose an approach to support knowledge mobilization and learning in networked innovation projects. Scenarios, stored in a shared repository, are used to capture and share information about application and solution domains. A collaborative process guides participants to reach a shared understanding and construct shared meaning. Stakeholders engage in a collaborative decision-making process of scenario ranking that includes identifying and negotiating comparison criteria. Although the approach is presented with examples in the domain of agriculture, where validation of the constituent elements took place, it is domain independent. - Documento de conferencia
Acceso Abierto Towards a Flexible and Scalable Data Stream Algorithm in FCA(2023) Leutwyler, Nicolás; Lezoche, Mario; Torres, Diego; Panetto, HervéThe amount of different environments where data can be exploited have increased partly because of the massive adoption of technologies such as microservices and distributed architectures. Accordingly, approaches to treat data are in constant improvement. An example of this is the Formal Concept Analysis framework that has seen an increase in the methods carried out to increment its capabilities in the mentioned environments. However, on top of the exponential nature of the output that the framework produces, the data stream processing environment still poses challenges regarding the flexibility in the usage of FCA and its extensions. Consequently, several approaches have been proposed to deal with them considering different constraints, such as receiving unsorted elements or unknown attributes. In this work, the notion of flexibly scalable for FCA distributed algorithms consuming data streams is defined. Additionally, the meaning of different scenarios of lattice merge in a particular data stream model is discussed. Finally, a pseudo-algorithm for merging lattices in the case of disjoint objects is presented. The presented work is a preliminary result and, in the future, it is expected to cover the other aspects of the problem with real data for validation.