PhD student

Rodrigo Gonzalez Laiz

I build world models that recover the hidden dynamics of complex systems, with a focus on neuroscience.

Animated trajectories of the Aizawa attractor.
Aizawa drag to rotate

About

Rodrigo Gonzalez Laiz

Hi! I'm a third-year PhD student at Helmholtz Munich, advised by Steffen Schneider.

My research explores how we can learn world models that capture the underlying dynamics of complex systems, particularly in scientific domains such as neuroscience. I develop self-supervised learning methods that recover meaningful and identifiable representations & dynamics from high-dimensional, nonlinear observations.

Long term, I want to bridge representation learning, dynamical systems and scientific discovery, so we can go from complex observations to models that help us understand, predict and control real-world systems.

Research

* equal contribution

  1. Figure from world2vec, showing multi-step dynamics in latent space. arXiv

    world2vec: Mechanistic world models through latent dynamics learning

    Tobias Schmidt*, Rodrigo Gonzalez Laiz*, Krisztián Szuppinger, Steffen Schneider
  2. Figure from Understanding the Limits of Latent World Models. arXiv

    Understanding the Limits of Latent World Models

    Rodrigo Gonzalez Laiz, Steffen Schneider
  3. Figure from Self-supervised contrastive learning performs non-linear system identification. ICLR 2025

    Self-supervised contrastive learning performs non-linear system identification

    Rodrigo Gonzalez Laiz*, Tobias Schmidt*, Steffen Schneider
  4. Figure from Time-series attribution maps with regularized contrastive learning. AISTATS 2025

    Time-series attribution maps with regularized contrastive learning

    Steffen Schneider, Rodrigo Gonzalez Laiz, Anastasiia Filippova, Markus Frey, Mackenzie Weygandt Mathis

Background

Experience

  1. 2024 – now

    PhD student Helmholtz Munich

    Identifiable representation learning, dynamical systems and world models, with applications to neuroscience.

  2. 2023

    Research Engineer EPFL, Mackenzie W. Mathis Lab

    Contributed to CEBRA, a software package for dimensionality reduction and representation learning of neuroscience data.

Education

  1. 2020 – 2022

    MSc Statistics ETH Zurich

  2. 2016 – 2020

    BSc Economics, Statistics & Computer Science Universidad Carlos III de Madrid

    Including an exchange year at the University of California, Davis (2018 – 2019).

Other

  1. 2026

    Machine Learning Summer School MLSS, Tübingen

  2. 2025

    Dynamic Brain Workshop Allen Institute, Seattle

    Project-based workshop on computational and systems neuroscience.