Two researchers from the Hertie AI Data Science department presented their posters at #NeurIPS2024 in Vancouver.

Attending NeurIPS is always an intense experience, but especially so this time with over 16000 people participating, making it the largest NeurIPS conference yet. 

We, PhD student Ifeoma Nwabufo and PostDoc Sebastian Damrich, spent six jam-packed days of thought-provoking keynotes with unique perspectives on the future of machine learning, informative tutorials on the latest method developments, targeted scientific workshops, and, most importantly, engaging poster sessions presenting the 4000 papers accepted this year. 

Sebastian Damrich presented his work Persistent homology for high-dimensional data based on spectral methods on detecting topological features in high-dimensional data, such as single-cell RNA-sequencing data. This attracted interest from both theoretical researchers in topological data science and applied computational biologists working with single-cell data. 

Ifeoma Nwabufo presented her work on Self-supervised visualisation of microscopy datasets at the Women in Machine Learning workshop. This work focused on using self-supervised learning to visualise medical images in two dimensions in order to gain meaningful insights about the structure of the data.

Besides the main program, the conference offered lots of opportunities for us to reconnect with researchers from all over the world and to get to know potential new collaborators during informal brainstorming sessions, exploring amazing Vancouver downtown, or at the ELLIS social.