Note that these are just example of research questions, and that I'm open to variations

This page maintains a list of possible internships.
For context, I'm associate professor at Lyon 1 University, LIRIS lab, in the DM2L team, Data Mining and Machine Learning.

I'm interested in the analysis of complex networks, and more generally in the analysis of complex systems. I'm also interested in the application of those methods to real world problems, in particular in the field of epidemiology, virology, cryptocurrencies, volcanic plumbing systems, science of science.

In 2023/2023, I have targeted fundings with co-supervision for 2 topics:

I'm interested in the analysis of complex networks, and more generally in the analysis of complex systems. I'm also interested in the application of those methods to real world problems, in particular in the field of epidemiology, virology, cryptocurrencies, volcanic plumbing systems, science of science.

In 2023/2023, I have targeted fundings with co-supervision for 2 topics:

- Graph Neural Networks with Attention to Understand the Colonization of Mosquito Larval Habitats According to their Biotic and Abiotic Characteristics
- Self-supervised learning for link prediction

Large language models are certainly one of the most important breakthrough in AI in the last decade. They are based on the idea that a model can be trained to predict the next word in a sentence, and that this model can then be used for a wide range of tasks. As for any deep neural network, the model is composed of *parameters*, or weights, which control the "neurons" of the neural network. The objective of this internship would be to start from a pretrained model such as llama or mistral, to download the weights of the model, and then to propose a modeling of this data in term of a network. The objective will therefore to use network science tools, such as community detection or centralities, to better understand the organization of such models, both "post mortems", i.e., the weights themselves, or in a "IRM" fashion, i.e., observing the evolution of the activity inside the "brain" when answering some specific requests. This is an exploratory subject, and the objective is to explore the potential of this approach, and to propose new methods to analyze those models.