PhD Program in Mathematics UNIFE-UNIMORE-UNIPR

Department

Department of Physics, Informatics and Mathematics,  University of Modena e  Reggio Emilia

Main research activities/topics/projects

The research project focuses on the intersection of Geometric Deep Learning and Computational Linguistics. The candidate will investigate the use of Graph Neural Networks (GNNs) to model and scale the Parametric Comparison Method (PCM).

Traditional linguistic analysis often treats syntactic features as independent; this project proposes to model languages as nodes in a graph where edges represent complex structural, genealogical, or geographical dependencies.

Key activities include:

- Developing GNN architectures (e.g., Graph Attention Networks) for missing data imputation in low-resource language datasets.

- Exploring Link Prediction to infer historical relationships between linguistic families.

- Applying Explainable AI (XAI) techniques to interpret the learned topological features of the syntactic manifold.

Working language 

The working language is English

Special entry requirements

Strong background in Mathematics and Computer Science. Proficiency in Python and Deep Learning frameworks (PyTorch, PyTorch Geometric).

 Basic knowledge or strong interest in Graph Theory and Natural Language Processing.

Duration in months (min-max)

Phd Sandwich: 6 - 12