Laurits Fredsgaard Larsen

About me

I’m a PhD candidate in Applied Mathematics and Computer Science at the Technical University of Denmark, specifically within the Cognitive Systems section. My research is part of the Bayesian neural networks for molecular discovery project.

I started my PhD, titled “Uncertainty Quantification for Graph Neural Networks”, in September 2023, and I’m supervised by Mikkel N. Schmidt (main supervisor) and Michael Riis Andersen (co-supervisor).

From October 2025 to March 2026, I am a visiting researcher at the National Institute of Informatics (NII) in Tokyo, hosted by Prof. Mahito Sugiyama.

Research interests

My research focuses on uncertainty quantification for graph neural networks (GNNs) applied to molecular discovery. I aim to develop reliable machine learning models that can accelerate the development of new medicines. I am broadly interested in:

• Efficient Deep Ensembles: Deep ensembles are highly effective for uncertainty quantification, but their computational cost can be prohibitive. How can we develop more efficient ensemble methods?
• Scalable and Performant Bayesian Neural Networks: Bayesian neural networks offer a principled approach to uncertainty estimation, but scaling them to large datasets and complex models remains a challenge. How can we improve their scalability and performance?
• Uncertainty Quantification in Graph Neural Networks: GNNs excel at capturing complex relationships, but how can we effectively quantify their uncertainty? Can we leverage their unique architectural properties to develop novel uncertainty estimation techniques?
• Learning with Limited Labeled Data: Labeled molecule data is often scarce. How can we leverage unlabeled data, related modalities (e.g., omics, images, text), and prior knowledge to improve model performance?
• Integrating Machine Learning with Experimentation: How can we effectively integrate machine learning models into the experimental workflow (“lab-in-a-loop”) to accelerate the discovery process?

Publications

• Semi-Supervised Learning for Molecular Graphs via Ensemble Consensus
  Rasmus H. Tirsgaard, Laurits Fredsgaard, Marisa Wodrich, Mikkel Jordahn, Mikkel N. Schmidt.
  NeurIPS 2025 Workshop: AI for Accelerated Materials Design (AI4Mat) (Accepted).

• On Joint Regularization and Calibration in Deep Ensembles
  Laurits Fredsgaard, Mikkel N. Schmidt
  Transactions on Machine Learning Research (TMLR), 2025. (Accepted).

• Capsid-like particles decorated with the SARS-CoV-2 receptor-binding domain elicit strong virus neutralization activity
  C. Fougeroux, …, Laurits Fredsgaard et al.
  Nature Communications, 2021.
  Publisher

• Head-to-head comparison of modular vaccines developed using different capsid virus-like particle backbones and antigen conjugation systems
  Laurits Fredsgaard et al.
  MDPI Vaccines, 2021.
  Publisher

Background

I have a dual background in life science and computer science, holding degrees in both Molecular Biomedicine and Machine Learning & Data Science from the University of Copenhagen.

My experimental research focused on virology and vaccine design. I studied the Hepatitis C Virus at Osaka University’s RIMD under Yoshiharu Matsuura (2017) and for my bachelor’s project at Hvidovre Hospital with Jens Bukh (2018). Subsequently, my master’s thesis (2018-2020) in Adam Sander’s lab centered on cVLP-based vaccines, where I supported the preclinical development of the ABNCoV2 COVID-19 vaccine, which has since proven highly successful in clinical trials.

My transition towards computational work includes my role as a student bioinformatician at LEO Pharma, where I built data analysis pipelines for omics data. For my final project in Machine Learning (2022-2023), I had the opportunity to work with Wouter Boomsma on deep learning models for protein sequences and structures.