David Grosser - Academia.edu (original) (raw)

Papers by David Grosser

HAL (Le Centre pour la Communication Scientifique Directe), Sep 20, 2010

This article argues that a cooperatively-built well-organized shared knowledge base is a new-and,... more This article argues that a cooperatively-built well-organized shared knowledge base is a new-and, from certain viewpoints, optimal-kind of support (refining and integrating other kinds of supports) for the following three complementary tasks: learning about living entities (and how to identify them), supporting their identification, and sharing knowledge about them. This article gives the ideas behind our prototype and argues that knowledge providers can be not solely specialists but also amateurs. In essence, for these three tasks, it argues for the (re-)use of much more semantically organized and interconnected versions of semantic wikis or scratchpads.

Predicting stability in object-oriented (OO) software, i.e., the ease with which a software item ... more Predicting stability in object-oriented (OO) software, i.e., the ease with which a software item evolves while preserving its design, is a key feature for software maintenance. In fact, a well designed OO software must be able to evolve without violating the compatibility among versions, provided that no major requirement reshuffling occurs. Stability, like most quality factors, is a complex phenomenon and its prediction is a real challenge. In this paper, we present an approach which relies on the case-based reasoning (CBR) paradigm and thus overcomes the handicap of insufficient theoretical knowledge on stability. The approach explores structural similarities between classes, expressed as software metrics, to guess their chances of becoming unstable. In addition, our stability model binds its value to the impact of changing requirements, i.e., the degree of class responsibilities increase between versions, quantified as the stress factor. As a result, the prediction mechanism favors the stability values for classes having strong structural analogies with a given test class as well as a similar stress impact. Our predictive model is applied on a testbed made up of the classes from four major version of the Java API.

HAL (Le Centre pour la Communication Scientifique Directe), 2007

Concept Lattices and their Applications, 2007

This paper presents an approach to Concept Analysis of structured, multivalued and incomplete dat... more This paper presents an approach to Concept Analysis of structured, multivalued and incomplete data currently present in life science knowledge bases. We are concerned with tree structured objects, whose size may be variable. We focus on the composition relations between attributes in the learning process. The interest of the method is the ability to take into account both structural and value parts of the objects. An application on a coral knowledge base illustrates the advantages of the method. 3 Knowledge representation model The knowledge representation model is made of the descriptive model and its instances, the structured objects.

2.1 La démarche expérimentale des systématiciens Tout au long de sa recherche, le systématicien a... more 2.1 La démarche expérimentale des systématiciens Tout au long de sa recherche, le systématicien apprend à reconnaître les espèces à partir d'un travail de bibliographie dans les muséums, d'observations effectuées en laboratoire sur des échantillons et d'échanges d'informations avec d'autres chercheurs. Il construit progressivement un modèle de description de son domaine qu'il applique à de nouvelles observations. Les erreurs d'identification l'amènent à remettre en cause son propre modèle. Il acquiert ainsi petit à petit une intuition du domaine, qui l'élève progressivement au rang d'expert reconnu par la communauté scientifique. Ce cheminement procède par un aller-retour, une itération entre les L'approche actuelle fondée uniquement sur l'informatisation des données de la systématique, généralement à l'aide des bases de données relationnelles est insuffisante. Bien sûr, celles-ci offrent de puissants moyens pour stocker un grand nombre d'informations et y accéder par l'intermédiaire de requêtes, mais elles présentent un certain nombre de limitations, comme nous l'avons souligné dans Rousse et al. (2000). 2.6 Limites des bases de données relationnelles A l'heure actuel, les Systèmes de Gestion de Bases de Données Relationnelles (SGBDR) sont largement utilisées pour assurer le stockage (persistance) et la diffusion des données. Les entités sont représentées par un ensemble de tuples, organisés en tables inter-reliées à l'aide d'index. Les tables peuvent être consultées ou générées dynamiquement (on appelle vues ces tables virtelles), par l'intermédiaire de requètes, généralement exprimées dans le langage standard SQL. 9. L'exposé des besoins précis de la Systématique en matière de représentation des connaissances, ainsi les différents modèles informatiques feront l'objet d'une étude détaillée au chapitre suivant. 10. Cf. annexe E pour une liste d'adresses de sites Internet concernés par le domaine.

HAL (Le Centre pour la Communication Scientifique Directe), May 14, 2008

HAL (Le Centre pour la Communication Scientifique Directe), 2008

One of the aims of expert knowledge management via Information and Communication Technology is to... more One of the aims of expert knowledge management via Information and Communication Technology is to improve the efficiency of knowledge transfer to non-specialists, and to facilitate the implementation of service-products that are adapted so as to be truly used. The Co-Design Platform (CDP) allows the service designers and users to determine service-product definitions together, to facilitate the emergence of their uses. Drawing on activity theory, this article describes a methodology that aims to guide the design process along the lines of the usage process. To illustrate this method, we've designed and implemented an instrumental learning e-service for guitar music (E-guitar). The Co-Design Platform gives a greater understanding of the transformation of the tool (proposals) into an instrument (proven demand), which is essential to the process of supplying the demand.

In the context of sustainable development of insular tropical islands, and more specifically for ... more In the context of sustainable development of insular tropical islands, and more specifically for sustainable education in the South West of Indian Ocean, data and knowledge management of specialists of natural or cultural diversity is at the heart of designing new ICT services. For example, the objectives of these e-services are to manage biodiversity and musical information on the Web, in order to preserve insular tropical islands common heritage. But the method of building data and knowledge bases is moving towards more Open, Inclusive and Smart approaches for a new 2020 Horizon. Open was initially inspired by EU (INSPIRE directive) and is characterized by opening public databases, for them to be enhanced by companies in new useful e-services for citizens. As such, Web Services are used for mutual inter-operability of databases. Inclusive is related to the different types of people that can participate to data and knowledge creation, i.e. experts, managers, stake- holders, amateur...

For future biodiversity studies relying on species identification, environmental officers and res... more For future biodiversity studies relying on species identification, environmental officers and researchers will only be left with monographic descriptions and collections in museums. This is why a knowledge base on the zooxanthellate scleractinian corals of the Mascarene Archipelago is being developed. This project offers results for both biologists/taxonomists and students or MPA-teams. Two online computer-based applications permit to identify genera and species. The first identification tool, called Xper2, was developed by LIS (Informatic and Systematics Laboratory) in Paris, and is used for identifications to genera. The second tool, named IKBS (Iterative Knowledge Base System), was developed by IREMIA (Institute for Research in Applied Mathematics and Computer Science) in La Reunion, and is used for identifications from families to species. The tools presently work for Astrocoeniidae, Pocilloporidae, Acroporidae (only Acropora + Isopora), Psammocoridae, Siderastreidae (owns Psamm...

Lecture Notes in Computer Science, 1999

In many fields dependant upon complex observation, the structuring, depiction and treatment of kn... more In many fields dependant upon complex observation, the structuring, depiction and treatment of knowledge can be of great complexity. For example in Systematics, the scientific discipline that investigates biodiversity , the descriptions of specimens are often highly structured (composite objects, taxonomic attributes), noisy (erroneous or unknown data), and polymorphous (variable or imprecise data). In this paper, we present IKBS, an Iterative Knowledge Base System for dealing with such complex phenomena. The originality of this system is to implement the scientific method in biology: experimenting (learning rules from examples) and testing (identifying new individuals, improving the initial model and descriptions). This methodology is applied in the following ways in IKBS: Knowledge is acquired through a descriptive model that suits the semantic demand of experts, 1. Knowledge is processed with an algorithm derived from C4.5 in order to take into account structured knowledge introduced in the previous descriptive model of the domain, 2. Knowledge is refined through the use of an iterative process to evaluate the robustness of the descriptive model and descriptions. 3. The IKBS system is presented here as a life science application facilitating the identification of coral specimens of the family Pocilloporidae.

BMC Ecology, 2013

Biodiversity informatics plays a central enabling role in the research community's efforts to add... more Biodiversity informatics plays a central enabling role in the research community's efforts to address scientific conservation and sustainability issues. Great strides have been made in the past decade establishing a framework for sharing data, where taxonomy and systematics has been perceived as the most prominent discipline involved. To some extent this is inevitable, given the use of species names as the pivot around which information is organised. To address the urgent questions around conservation, land-use, environmental change, sustainability, food security and ecosystem services that are facing Governments worldwide, we need to understand how the ecosystem works. So, we need a systems approach to understanding biodiversity that moves significantly beyond taxonomy and species observations. Such an approach needs to look at the whole system to address species interactions, both with their environment and with other species. It is clear that some barriers to progress are sociological, basically persuading people to use the technological solutions that are already available. This is best addressed by developing more effective systems that deliver immediate benefit to the user, hiding the majority of the technology behind simple user interfaces. An infrastructure should be a space in which activities take place and, as such, should be effectively invisible. This community consultation paper positions the role of biodiversity informatics, for the next decade, presenting the actions needed to link the various biodiversity infrastructures invisibly and to facilitate understanding that can support both business and policy-makers. The community considers the goal in biodiversity informatics to be full integration of the biodiversity research community, including citizens' science, through a commonly-shared, sustainable e-infrastructure across all sub-disciplines that reliably serves science and society alike.

A new iterative and interactive algorithm called CSN (Classification by Successive Neighborhood) ... more A new iterative and interactive algorithm called CSN (Classification by Successive Neighborhood) to be used in a complex descriptive objects identification approach is presented. Complex objects are those designed by experts within a knowledge base to describe taxa (monography species) and also real organisms (collection specimens). The algorithm consists of neighborhoods computations from an incremental basis of characters using a dissimilarity function which takes into account structures and values of the objects. A discriminant power function is combined with domain knowledge on the features set at each iteration. It is shown that CSN consistently outperforms methods such as identification trees and simplifies interactive classification processes comparatively to search for K-Nearest-Neighbors method. Index Terms — identification, Similarity, K-Nearest-Neighbors, Decision Trees, structured data, knowledge base, life science. —————————— u ——————————

Formal concept analysis (FCA) has proved helpful in the resolution of practical problems from fie... more Formal concept analysis (FCA) has proved helpful in the resolution of practical problems from fields such software engineering, knowledge engineering and data mining. Recently, a substantial push has been done toward the design of efficient procedures for lattice construction, with a variety of novel algorithms proposed in the literature. However, the FCA community has created only few effective tools for manipulating lattices so far and what is still missing is an integrated environment for constructing, visualizing, exploring and maintaining lattices. We present the Galicia project aimed at the construction of an open platform for lattice manipulation which follows the complete life-cycle of a lattice. More than just a lattice tool, the platform provides the necessary services for quick development and test of new lattice algorithms.

Formalization of scientific knowledge in life sciences by experts in biology or Systematics produ... more Formalization of scientific knowledge in life sciences by experts in biology or Systematics produces arborescent representations whose values could be present, absent or unknown. To improve the robustness of the classification process of those complex objects, often partially described, we propose a new classification method which is iterative, interactive and semi-directed. It combines inductive techniques for the choice of discriminating variables and search for nearest neighbors based on various similarity measures which take into account structures and values of the objects for the neighborhood computation.

Knowledge discovery from data in environmental sciences is becoming more and more important nowad... more Knowledge discovery from data in environmental sciences is becoming more and more important nowadays because of the deluge of information found in databases of digital ecosystems, coming altogether from institutions and amateurs. For example in biodiversity science, all these data need to be validated by specialists with the help of Intelligent Environmental Decision Support Systems (IEDSSs), then enhanced and certified into qualitative knowledge before reaching their audience. Data mining through classification or clustering is the dedicated inductive process of grouping descriptions based on similarity measures, then building classes and naming them. Later, the formed concepts can be reused for identification purpose with new observations. The problem is that when using such knowledge-based systems, we tend to forget the fundamental role of subjects (endusers) in the definition, observation and description of objects. In order to get good identification results, a consensus must b...

IFIP Advances in Information and Communication Technology, 2016

1 When using Artificial Intelligence techniques for Knowledge management in decision support syst... more 1 When using Artificial Intelligence techniques for Knowledge management in decision support systems, the enhancement of knowledge should be based both on artificial machine learning methods and a natural human learning approach. Indeed, knowledge representation with ontologies and Case-Based Reasoning (CBR) is not enough for gaining qualitative results in decision support systems. We need to manage know-how, i.e. living knowledge. For example, enhancing biodiversity and music means teaching and learning the effectiveness of individual and living interpretations (how to observe a natural specimen, how to play a music sheet). The quality of descriptions is thus very important to correctly classify or identify marine or terrestrial organisms, or learn adequately an instrument such as the guitar or the piano. This paper introduces Sign management to tackle this qualitative learning problem in AI. Then, a Codesign methodology and a cooking method on a Creativity Platform are proposed: when dealing with such complex domains, we need to focus on the signification of knowledge construction that operates in co-designing an eservice that should be useful for reaching a more robust knowledge base. Our finding is that due to different interpretations of domain objects from subjects (persons), we need sign bases to move from written expert knowledge transmission to multimedia know-how sharing in the community for getting better results.