Skip to content
View lenz3000's full-sized avatar
11 followers · 3 following
  • TU Berlin
  • Berlin

Achievements

Achievement: Pull Sharkx2Achievement: QuickdrawAchievement: Arctic Code Vault Contributor

Achievements

Achievement: Pull Sharkx2Achievement: QuickdrawAchievement: Arctic Code Vault Contributor

Block or report lenz3000

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
lenz3000/README.md

Hi! I'm Lorenz Vaitl, a PhD graduate in Machine Learning from Technische Universität Berlin (2024). Over the past few years, I've been deeply involved in researching and teaching, focusing on generative models and their applications to Quantum Field Theory and Variational Inference.

🚀 About Me

  • 🌟 Primary Interests:

    • Generative Models
    • Normalizing Flows
    • Monte Carlo Methods
    • Variational Inference
    • Bayesian Statistics

  • 🎓 Education:

    • Ph.D. in Machine Learning, 2024
      Technische Universität Berlin
    • M.Sc. in Computer Science, 2019
      Technische Universität Berlin
    • B. Sc. in Computer Science, 2015
      RWTH Aachen

🔬 Research Highlights

  • Learning trivializing gradient flows for lattice gauge theories:
    We derived a highly efficient Continuous Normalizing Flow for simulating Lattice Gauge Theories. The Model contains all the symmetries of the Theory and set a new state of the art.

  • Path Gradients: The major focus of my research was on deriving low-variance gradient estimators for Normalizing Flows. We published three papers at ICML 2022(oral), ICLR 2024, and in MLST, with every contribution covering a different aspect from generalizations of path gradients, thorough analysis and efficient algorithms for various architectures and loss functions. For the full picture check out my thesis

🛠️ Skills

  • Programming Languages: Python, Latex, C++
  • Libraries & Frameworks: PyTorch
  • Tools: Git, neovim, SLURM, ssh
  • Core Areas: Generative Models, Bayesian Inference, Machine Learning, Deep Learning

📫 Let's Connect!

Feel free to explore my repositories for more insights into my research and coding projects.

Popular repositories Loading

  1. ffjord-path ffjord-path Public

    Forked from rtqichen/ffjord

    code for "Path-Gradient Estimators for Continuous Normalizing Flows".

    Python 5 2

  2. unified-path-gradients unified-path-gradients Public

    Jupyter Notebook 4

  3. DoG-CNNs DoG-CNNs Public

    Difference of Gaussian filters as Convolutional Filters

    Jupyter Notebook 2 1

  4. path-grads-after-fm path-grads-after-fm Public

    Jupyter Notebook 2

  5. Sparse-Logit-GP-LMM Sparse-Logit-GP-LMM Public

    This was my Master's Thesis. I extended the work on Efficient Gaussian Process Classification to a LMM, that was a linear classifier that models confounders via a SGP

    Python

  6. lenz3000 lenz3000 Public