Research Internship proposal

Internship: Large Language Models for Knowledge Graphs Validation

Duration: 6 months

Location: IRIT, 118 route de Narbonne, Toulouse, France

Advisors:
Farah Benamara,
IRIT, Toulouse University and IPAL-CNRS, Singpoure
farah.benamara@irit.fr

Gratification: 630¤/mouthn

Context

The Institut de Recherche en informatique de Toulouse (IRIT-Toulouse
University) has a 6 months position oOer within the ANR PRCI project
QualityOnt
(https://helios2.mi.parisdescartes.fr/~sseifedd/qualityont/) on using
NLP for knowledge graph validation from textual data (social media and
news) in a multilingual context.

The project aims at generating high-quality knowledge graph (KG) for
emergent English, French and German trends during sanitary crises
situations through (1) a general view of how facts about the pandemic
evolve across time and languages, and (2) a high-quality evaluation of
these facts in enriched knowledge graphs to support further
analysis. The methodology and results of QualityOnt were designed to
be generic enough to ensure their reusability in other future sanitary
crises situations.

Objectives

The position is funded by an ANR PRCI-Project involving academics from
NLP (Toulouse Univ.- IRIT), Database quality (Univ. de Paris-LIPAD)
and semantic web/knowledge graph (Univ.  Lubeck-Germany). The
candidate will start from a knowledge graph (KG) already developed by
within the project will be in charge of developing NLP techniques for
KG validation, in particular: (1) Write a state of the art on
extracting triples (facts) and validate them from either unstructured
or semi-structured data, (2) Identify existing tools/methods and
re-implement them (to be used as baselines), (3) Propose new solutions
based on LLMs to account for negation and modality at the sentence
level.

References:

[1] Morteza Kamaladdini Ezzabady, Frédéric Ieng, Hanieh
Khorashadizadeh, Farah Benamara, Sven Groppe, Soror Sahri: Towards
Generating High-Quality Knowledge Graphs by Leveraging Large Language
Models. NLDB (1) 2024: 455-469

[2] Hanieh Khorashadizadeh, Fatima Zahra Amara, Morteza Kamaladdini
Ezzabady, Frédéric Ieng, Sanju Tiwari, Nandana Mihindukulasooriya,
Jinghua Groppe, Soror Sahri, Farah Benamara, Sven Groppe: Research
Trends for the Interplay between Large Language Models and Knowledge
Graphs. CoRR abs/2406.08223 (2024)


To apply, send your CV+grades (relevés de notes) to
farah.benamara@irit.fr by December 20th 2024