Offre en lien avec l’Action/le Réseau : – — –/– — –

Laboratoire/Entreprise : Le Laboratoire d’Informatique, de Robotique et de
Durée : 6 mois (+ CDD 2 mois
Contact : alexandre.bazin@lirmm.fr
Date limite de publication : 2023-12-15

Contexte :
To ensure the success of the agroecological transition, farmers need to have access to knowledge about alternatives to conventional farming techniques. However, before a knowledge base (KB) can be used by farmers and scientific experts, it needs to be corrected of its anomalies.

Sujet :
The context of this internship is the Knomana KB [Silvie et al., 2021], which brings together 48,000 descriptions of pesticidal and antibiotic uses of plants, and aims to propose plant-based preparations to replace synthetic chemicals. Dictionaries are already available to correct values for its 31 data types. But, verifying data correction and consistency is too complex to be carried out manually. For example, an inconsistency between the pesticide plant, the protected system (e.g. corn crop), the bio-aggressor (e.g. insect) and the geographical location is enough to mislead a farmer. The method named Attribute Exploration (EA), developed by Formal Concept Analysis, can be used to detect and correct these anomalies [Saab et al., 2022]. EA expresses each piece of knowledge in the form of an implication rule, and identifies generalizations at different levels (e.g. all insects of genus X are controlled by plants of Family Y). The rules are presented to the experts, who validate or invalidate them in order to bring the BC into a coherent state.
The objective of the internship is to develop a software prototype for detecting and correcting anomalies in multidimensional and multirelational data. This prototype will enable to manipulate data and data types, then interact with the FCA4J library, for rule computation, and the RCAvizIR software, developed with the support of #Digitag (Master internships in 2022 and 2023) to present them in an order that facilitates correction work by experts.
The work will be conducted according to the design stages described by Sedlmair et al. 2012 (literature study, definition of the need in terms of a visual problem, proposal of a mock-up, development, deployment, validation).

* Michael Sedlmair, Miriah D. Meyer et Tamara Munzner. Design Study Methodology: Reflections from the Trenches and the Stacks. IEEE TVCG 18(12): 2431-2440, 2012.

Profil du candidat :
Student in Master studies (computer science or bioinformatics)

Formation et compétences requises :
Strong skills in programming and data analysis, with an interest for knowledge engineering, visual analytics, and to find alternatives to chemical pesticides and antibiotics in organic agriculture.

Adresse d’emploi :
The student will be integrated in the Web3 teams of LIRMM, in Montpellier, and will collaborate with researchers from Advance and Marel teams.