Vincent Jung

News

Research

I work on developing AI models that are designed to extract biologically relevant features from large datasets. Currently, I am working with RNA Language Models, which are trained similarly to LLMs but on RNA sequences. I seek to understand what these models fundamentally capture, how they do so and how we can improve existing model designs or create new architectures or training paradigms altogether. Moreover, I work on their interpretation and application to relevant biological issues such as structure or interaction prediction.

This field of research is particularly exciting to me as it raises many important question in deep learning: how do we design models meant to capture the molecular world, how do we include prior knowledge while letting models learn their own, how do we efficiently learn on large databases with many redudancies, and so on. It also deals with real biological problems, and improvements in our application tasks can quickly lead to unlocking new possibilities in biological research.

Although I have a computer science background, I work closely with RNA biologists and aim to understand the underlying biology of the subjects I work on in depth. In this, I am lucky to be part of a now long-running collaboration between my three supervisors each coming with their own expertise, namely AI, RNA Biology and NLP. This collaboration has already resulted in a Nature Reviews Molecular Cell Biology paper which outlines how these three subjects can come together to unlock new possibilities in RNA Biology.

Publications

Semester and Master projects