PhD Position F / M The role of rapport in humanconversational agent interaction : Modeling conv[...]

INRIA; Paris; Sur place; EUR 30 000 - 35 000; Plein temps

Résumé du poste

Un institut de recherche en informatique à Paris recherche un doctorant pour développer des agents conversationnels. Le candidat devra analyser le comportement conversationnel et améliorer les capacités sociales des modèles de langage. Des compétences solides en apprentissage profond, programmation Python, ainsi qu'une bonne maîtrise du français et de l’anglais sont requises. Des avantages intéressants comme des repas subventionnés et la possibilité de télétravail sont offerts.

Prestations

Repas subventionnés

Remboursement partiel des transports

7 semaines de congés annuels + 10 jours RTT

Télétravail possible et horaires flexibles

Accès à des équipements professionnels

Avantages sociaux, culturels et sportifs

Formation professionnelle

Sécurité sociale

Qualifications

• Bonne compétence en apprentissage profond et en statistiques.
• Expérience en systèmes de dialogue appréciée.
• Maîtrise du français et de l'anglais.

Responsabilités

• Concevoir et conduire des expériences.
• Analyser les résultats.
• Écrire une thèse de doctorat.

Connaissances

Compétence en apprentissage profond appliqué au langage

Connaissance des systèmes de dialogue

Compétences avancées en programmation Python

Compétences en statistiques

Compétences en C++

Formation

Doctorat en intelligence artificielle ou domaine connexe

Outils

TensorFlow

PyTorch

Description du poste

Contexte et atouts du poste

The objective of this project is to build embodied conversational agents (also known as ECAs, virtual humans, chatbots, or multimodal dialogue systems) capable of engaging users in both social and task-oriented conversations, where social talk enhances task performance. To achieve this, we model human-human conversation, integrate these models into ECAs, and evaluate their effectiveness.

This project is based at the prestigious INRIA computer science institute in downtown Paris, within the PRAIRIE Institute for Interdisciplinary Research on AI, one of the four 3IA institutes launched by the French government in 2019. For more information, see:

Mission confiée

The doctoral student selected will develop skills in analyzing and synthesizing conversational behavior. They will explore ways to improve the social capabilities of Large Language Models (LLMs) by developing methods to build and maintain rapport in human-AI interactions. The research includes building and annotating a corpus focused on social aspects of human conversation within a specific domain, analyzing data to derive computational models informed by empirical findings and literature, and refining these models for integration into LLM-based conversational systems. The goal is to preserve task performance while enabling socially aware behavior. The models will be integrated into an Embodied Conversational Agent (ECA) to evaluate their impact on system performance. Potential applications include digital assistants, educational tools, and healthcare AI systems.

Principales activités

• Design and conduct experiments
• Analyze results
• Implement modules in socially-aware dialogue systems
• Write articles
• Present work at conferences
• Write a doctoral thesis

Compétences

Technical skills and level required: Solid competence in deep learning applied to language (experience with dialogue systems is a strong plus), proficiency in statistics, advanced programming skills in Python and C++, and familiarity with tools such as TensorFlow and PyTorch.

Languages: French and English

Relational skills: Ability to work in a team and collaborate across disciplines and backgrounds.

Other valued skills: Theoretical background in cognitive science, linguistics, conversational analysis, sociolinguistics, social cognition, or learning science.

Avantages

• Subsidized meals
• Partial reimbursement of public transport
• 7 weeks of annual leave + 10 RTT days (full-time basis) + possibility of exceptional leave (e.g., sick children, moving)
• Possibility of telework and flexible working hours
• Access to professional equipment (video conferencing, IT equipment loans, etc.)
• Social, cultural, and sports benefits (Inria social works association)
• Access to professional training
• Social security