Group Burkholz has several open PhD positions. Engagement in topics related to deep learning efficiency in association with an ERC Starting Grant (SPARSE-ML) and a collaboration with Apple Research on LLM efficiency would be a plus.

About the project:

Possible research topics include: 

• discrete probabilistic models
• robust learning algorithms
• meaningful applications
• molecular biology and genetics (in collaboration with domain experts at Harvard)
• international food trade
• drug discovery and more

• deep learning (theory and applications)
• machine learning efficiency
• sparse learning and modelling (e.g. lottery tickets)
• gradient compression
• graph neural networks and complex networks

You can check out Rebekka's website to know more about her!

"Deep learning continues to achieve impressive breakthroughs across disciplines and is a major driving force behind a multitude of innovations. For instance, it has helped computers convert text into images, assist in the design of new drugs, generate code, and even find fast matrix multiplication algorithms. Yet, it comes at considerable costs. It requires huge amounts of data to learn complex models that are difficult to interpret and are vulnerable to attacks.

The major goal of the Relational Machine Learning Lab is to alleviate these problems based on sound theoretical insights and broaden the applicability of deep learning across the natural sciences and biomedical domain."