Missions

The candidate will work on neuro-symbolic Artificial Intelligence as part of the ANR JCJC ERIANA project. Commonsense knowledge plays an increasingly important role in the development of AI systems. Knowledge can be expressed using different representation formalisms including symbolic frameworks and vector representations where each of these formalisms has complementary strengths and weaknesses when it comes to implementing reasoning. The postdoc positions will focus on developing neuro-symbolic methods for learning and reasoning while ensuring transparency and explainability of obtained results. A particular focus will be on learning concept/relation embeddings and event-centric language understanding.

Activités

The successful candidate will work in collaboration with members of the ANR JCJC ERIANA project to develop a framework for the neuro-symbolic representation of narratives that allows events and their interactions to be encoded in an interpretable way. Particular emphasis will be placed on the learning of concept and relation embeddings, and the use of these embeddings to implement reasoning tasks. The successful candidate will then develop deep generative models leading to the discovery of knowledge that can be considered as symbolic rules and combined to perform grounded reasoning.

Compétences

- PhD in Artificial Intelligence with skills in deep learning, natural language processing and knowledge representation and reasoning.
- Proficiency in Python and the deep learning platforms Pytorch and Tensorflow.
- Fluency in written and spoken English (no knowledge of French required).

Contexte de travail

The successful candidate will work at the Centre de Recherche en Informatique de Lens (CRIL UMR 8188), a laboratory of the Université d'Artois and the CNRS whose research focuses on artificial intelligence and its applications. CRIL has over 60 members and is a member of the CLAIRE European Confederation of Artificial Intelligence Laboratories and the humAIn regional alliance.