Lars Doorenbos

profile.jpeg

University of Bern

Bern, Switzerland

I am a Ph.D student currently based in Bern, Switzerland, where I work on creating safe and reliable deep learning models through out-of-distribution detection. Before, I completed my Bachelor’s and Master’s degree in Computing Science at the University of Groningen.

news

Nov 15, 2024 I completed my PhD!
Oct 28, 2024 Our work Galaxy spectroscopy without spectra: Galaxy properties from photometric images with conditional diffusion models was accepted at The Astrophysical Journal.
Jul 10, 2024 Our work on predicting galaxy spectra from photometry was accepted as an oral at the International Conference on Machine Learning for Astrophysics. I will also give a flash talk on ULISSE.
Jul 1, 2024 Our paper “Learning Non-Linear Invariants for Unsupervised Out-of-Distribution Detection” has been accepted at the European Conference on Computer Vision.
Jun 30, 2024 Our paper “SuFIA: Language-Guided Augmented Dexterity for Robotic Surgical Assistants” has been accepted at the International Conference on Intelligent Robots and Systems. It was also featured in a New Scientist article and the C4SR+ workshop at ICRA 2024.
May 21, 2024 Our paper “Hyperbolic Random Forests” has been published in Transactions on Machine Learning Research.
Apr 2, 2024 I have been selected as a DAAD AInet fellow for 2024.
Nov 10, 2023 I finished my internship at Nvidia, where I worked on controlling robots with vision-language models. A preprint of my work is currently under review.
Jul 18, 2023 Our paper “Stochastic Segmentation with Conditional Categorical Diffusion Models” has been accepted at ICCV 2023.
Apr 20, 2023 “Generating astronomical spectra from photometry with conditional diffusion models” has been selected as an oral presentation at the Bern Data Science Day 2023.

selected publications

2024

  1. Learning Non-Linear Invariants for Unsupervised Out-of-Distribution Detection
    Lars Doorenbos, Raphael Sznitman, and Pablo Márquez-Neila
    European Conference on Computer Vision, 2024
  2. Galaxy spectroscopy without spectra: Galaxy properties from photometric images with conditional diffusion models
    Lars Doorenbos*, Eva Sextl*, Kevin Heng, and 6 more authors
    arXiv preprint arXiv:2406.18175, 2024
    *equal contribution
  3. Hyperbolic Random Forests
    Lars Doorenbos, Pablo Márquez-Neila, Raphael Sznitman, and 1 more author
    Transactions on Machine Learning Research, 2024

2023

  1. Stochastic Segmentation with Conditional Categorical Diffusion Models
    Lukas Zbinden*, Lars Doorenbos*, Theodoros Pissas, and 3 more authors
    International Conference on Computer Vision, 2023
    *equal contribution

2022

  1. Generating astronomical spectra from photometry with conditional diffusion models
    Lars Doorenbos, Stefano Cavuoti, Giuseppe Longo, and 3 more authors
    NeurIPS: Machine Learning for the Physical Sciences Workshop, 2022
  2. Data invariants to understand unsupervised out-of-distribution detection
    Lars Doorenbos, Raphael Sznitman, and Pablo Márquez-Neila
    In European Conference on Computer Vision, 2022