Analogical reasoning for KG management tasks : Entity Alignment and GraphRAG

INRIA; Valbonne; Sur place; EUR 20 000 - 40 000; Plein temps

Résumé du poste

Un institut de recherche renommé recherche un stagiaire pour étudier le raisonnement analogique sur des tâches de gestion de graphes de connaissances à Valbonne. Le candidat idéal aura une formation en informatique ou en mathématiques appliquées, maîtrisera Python et possédera des compétences en Machine Learning. Ce stage offre des avantages tels que des repas subventionnés, un remboursement partiel des frais de transport, et une possibilité de télétravail après 6 mois.

Prestations

Repas subventionnés

Remboursement partiel des frais de transport

7 semaines de congés annuels + 10 jours RTT

Travail flexible après 6 mois

Équipements professionnels fournis

Qualifications

• Étudiant(e) en informatique ou mathématiques appliquées.
• Maîtrise de Python et des concepts de Machine Learning.
• Curieux(se) et désireux(se) d'apprendre.

Responsabilités

• Étudier l'application du raisonnement analogique sur la gestion des KG.
• Appliquer le raisonnement analogique à l'alignement d'entités.
• Concevoir le pipeline expérimental pour RAG et GraphRAG.

Connaissances

Programmation en Python

Machine Learning / Deep Learning

Connaissance des LLMs

Connaissance du Web Sémantique (RDF, RDFS, OWL, SPARQL)

Anglais (capacité à lire et écrire)

Formation

Master Année 2 ou dernière année d'école d'ingénieur

Outils

PyTorch

TensorFlow

Description du poste

Contexte et atouts du poste

Analogical reasoning, expressed with analogical quadruples of the form “a is to b as c is to d” (e.g. Paris is to France as Berlin is to Germany), is a natural way for human beings to reason about new situations based on the knowledge gained from experiencing similar situations. Its insights have been proven in various human cognitive tasks, such as natural language learning or problem‑solving, and recently in machine learning through analogy‑based classifiers [9] and retrievers [5].

The past work of [6] has demonstrated that analogy‑based classifiers can be applied to knowledge graph (KG) management tasks, showing great results for domain‑specific KG bootstrapping, and paving the road for testing this analogy‑based classifier on other KG management tasks. Among those, we aim at studying:

• Entity alignment: where similar entities across different knowledge graphs should be detected [7,8] e.g. two entities representing the same city but in two different graphs.
• GraphRAG: where facts from KGs are used to ground LLMs [1,2,3,4] e.g. a graph representing the products of a company is used in answering questions about these products.

Mission confiée

In this internship, we propose to study the application of analogical reasoning on two KG management tasks.

This internship will take place on the premises of the Wimmics team in Sophia Antipolis, under the supervision of:

• Pierre Monnin
• Fabien Gandon
• Ndeye‑Emilie Mbengue

Wimmics (Web‑Instrumented huMan‑Machine Interactions, Communities and Semantics) is a joint research team at Université Côte d’Azur, Inria, CNRS, I3S, whose research lies at the intersection of artificial intelligence and the Web. Wimmics members work on methods to extract, control, query, validate, infer, explain and interact with knowledge.

Principales activités

In this internship, we propose to study the application of analogical reasoning on two KG management tasks. In particular, the internship will include the following tasks:

• Understanding key concepts of KG, entity alignment, GraphRAG, and analogical reasoning through a literature review.
• Applying analogical reasoning on entity alignment.
• Identifying benchmark datasets (with a potential extension to the Ontology Matching task).
• Designing the experimental pipeline.
• Implementation, experimentation and evaluation.
• Applying analogical reasoning on RAG / GraphRAG.
• Identifying RAG / GraphRAG components that could be replaced / enriched with analogical reasoning.
• Identifying benchmark datasets.
• Designing the experimental pipeline: for RAG and GraphRAG.
• Implementation, experimentation and evaluation.

Compétences

You are studying in Master Year 2 / final year of engineering school, with a specialty in computer science or applied mathematics. You are proficient in:

• Python programming.
• Machine Learning / Deep Learning, especially with frameworks like PyTorch or TensorFlow.
• Knowledge of LLMs, frameworks like LangChain, and (Graph)RAG would be appreciated.
• Knowledge of the Semantic Web (RDF, RDFS, OWL, SPARQL, knowledge graphs, and ontologies) would be appreciated.
• Ability to read and write in English.

You are curious, eager to learn, face challenges, experiment, and discover by yourself.

Avantages

• Subsidised meals.
• Partial reimbursement of public transport costs.
• Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.).
• Possibility of teleworking (after 6 months of employment) and flexible organisation of working hours.
• Professional equipment available (videoconferencing, loan of computer equipment, etc.).
• Social, cultural and sports events and activities.
• Access to vocational training.
• Social security coverage.