PhD Position F/M Neural Representations of Multilingual Language Processing

Contract type : Fixed-term contract

Level of qualifications required : Graduate degree or equivalent

Fonction : PhD Position

Context

The PhD will be supervised by Benoît Sagot at Inria (Inria Paris centre) in the ALMAnaCH project-team (http://almanach.inria.fr/index-en.html) and within the PRAIRIE-PSAI institute. It will be co-supervised by Demian Wassermann (Inria Saclay center, project-team MIND), and will take place in close collaboration with Christophe Pallier (NeurSpin). It will be financed by Benoît Sagot's PRAIRIE-PSAI chair.

Assignment

PhD topic

Understanding how the brain encodes language remains one of the core challenges in cognitive neuroscience. In recent years, advances in nat- ural language processing (NLP)—particularly with brain-inspired neural network architectures—have opened new perspectives to explore how linguistic information is represented. By bringing together tools from both neuroscience and NLP, researchers can now directly compare the represen- tations learned by computational models with patterns of brain activity. This cross-disciplinary approach enables the development of more grounded and data-driven hypotheses about how language is processed in the human brain.

Although some multilingual datasets include stimuli and recordings in multiple languages (Li et al., 2022), to our knowledge, there is no fMRI neuroimaging dataset in which bilingual individuals perform the same task in both of their languages. To address this, our plan is to collect fMRI data from bilingual participants who speak both their first language (L1) and second language (L2) at a native or near-native level. We will examine how their brains respond to each language and how those responses relate to representations generated by computational models.

Overall, this PhD project will use tools from NLP to better understand how the brain understands and organizes multiple languages. By comparing brain activity recorded during natural listening with deep learning model embeddings, we hope to reveal where artificial and biological language systems might converge. Focusing on bilingual individuals will allow us to study how the brain responds to the same meaning expressed in different languages. With this project, our aim is to understand whether the brain uses a shared or language-specific system to represent meaning and how closely these systems resemble the ones used by modern language models, or vice versa.

Main activities

Main activities

The candidate’s main activities will include:

  • keeping up-to-date with related work on the topic with regular reading
  • carrying out research on the topic outlined above, both in the development of new ideas, positioning with respect to related work and validation of the methodology via experiments and analysis
  • presenting their work work both internally to colleagues and externally in the form of conference/journal/workshop papers and in the final PhD thesis
  • interacting and exchanging with colleagues on related topics

The PhD position is a 3-year funded position to start from the 1st September 2025.

Skills

They should have a good level in programming (python), experience with neural networks and neuroscience and an interest in natural language processing and neuroimagery. A good written and spoken level of English is required. Knowledge of French and/or other languages is a plus.

Benefits package

  • Subsidized 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 and flexible organization 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