The BamlerLab reading group meets weekly on Thursdays to engage in a comprehensive exploration and interpretation of scholarly works. Your participation in the group is cordially invited, and should you choose to attend, please feel free to join in the discussion.

Additionally, we’re thrilled to introduce a new reading group exclusively centered on compression topics (). This exciting venture convenes every Monday. It promises to be an engaging forum for those passionate about compression algorithms and techniques.

We thank ChatGPT for writing this description.

Location and Time

  • MvL6: 4th floor seminar room
  • : Thursday 1:30 pm
  • : Monday 3:00 pm

Schedule

2024

Date Moderator Title of Paper & Link to Paper talks.tue comment reading group
2024/10/21 all participants Entropy-Constrained Training of Deep Neural Networks talks.tue  
2024/10/10 Robert Bamler Tutorial on Information Theory talks.tue  
2024/09/12 all participants Experiences and Trends at ICML 2024 N/A  
2024/08/08 Robert Bamler Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution talks.tue  
2024/07/29 all participants Neural Discrete Representation Learning talks.tue  
2024/07/15 all participants The No Free Lunch Theorem, Kolmogorov Complexity, and the Role of Inductive Biases in Machine Learning talks.tue  
2024/07/11 Tristan Cinquin Randomized Algorithms for Matrix Computations (Chapters 3-5) talks.tue  
2024/07/04 Robert Bamler Randomized Algorithms for Matrix Computations (Chapters 1-3) talks.tue  
2024/06/24 all participants Distribution Compression in Near-linear Time talks.tue  
2024/06/17 all participants Estimating the Rate-Distortion Function by Wasserstein Gradient Descent talks.tue  
2024/05/13 all participants Lossy Image Compression with Conditional Diffusion Models talks.tue  
2024/05/06 all participants Out-of-Distribution Detection using Maximum Entropy Coding talks.tue  
2024/04/29 all participants On universal quantization talks.tue  
2024/04/22 all participants Universal Deep Neural Network Compression talks.tue  
2024/04/11 Alexander Conzelmann Intrinsic Dimensionality Explains the Effectiveness of Language Model Fine-Tuning talks.tue  
2024/04/08 all participants Lossy Compression with Gaussian Diffusion talks.tue Part 2/2
2024/04/04 Tim Xiao LoRA: Low-Rank Adaptation of Large Language Models talks.tue  
2024/03/21 Tristan Cinquin Bayesian Model Selection, the Marginal Likelihood, and Generalization talks.tue  
2024/03/18 all participants Lossy Compression with Gaussian Diffusion talks.tue Part 1/2
2024/03/14 Johannes Zenn Diffusion Schrödinger Bridge Matching talks.tue Part 2/2
2024/03/11 all participants Language Modeling is Compression talks.tue  
2024/03/07 Johannes Zenn Diffusion Schrödinger Bridge Matching talks.tue Part 1/2
2024/03/04 all participants Bit-Swap: Recursive Bits-Back Coding for Lossless Compression with Hierarchical Latent Variables talks.tue  
2024/02/29 Alexander Conzelmann Learning Generative Models with Sinkhorn Divergences talks.tue  
2024/02/26 all participants Wasserstein Distortion: Unifying Fidelity and Realism N/A  
2024/02/19 all participants Compressing Images by Encoding Their Latent Representations with Relative Entropy Coding N/A  
2024/02/15 Robert Bamler Introduction to Optimal Transport N/A  
2024/02/08 Alexander Conzelmann Low-Precision Stochastic Gradient Langevin Dynamics N/A  
2024/01/11 Tim Xiao
&
Johannes Zenn		 Experiences and Trends at NeurIPS 2023 N/A  

2023

Date Moderator Title of Paper & Link to Paper talks.tue comment reading group
2023/12/07 Robert Bamler Show and Tell Session talks.tue Show and Tell Session
2023/11/30 Lenard Rommel Finite Volume Neural Networks for Simple Vortex Problems talks.tue Bachelor’s Thesis Presentation
2023/11/23 Johannes Zenn More Faithful Variational Inference via the Initial Distribution of Differentiable Annealed Importance Sampling talks.tue  
2023/10/26 Tristan Cinquin Regularized KL-Divergence for Well-Defined Function Space Variational Inference in BNNs talks.tue  
2023/09/07 Alexander Ludwig Neural Data Compression for Magnetic Resonance Imaging talks.tue Bachelor’s Thesis Presentation
2023/08/31 Nicolò Zottino Probabilistic Circuits talks.tue Talk
2023/08/17 Robert Bamler Algorithms for the Communication of Samples talks.tue  
2023/08/10 Robert Bamler Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow talks.tue  
2023/07/20 Tim Xiao DreamFusion: Text-to-3D using 2D Diffusion talks.tue  
2023/07/06 Alexander Conzelmann From data to functa: Your data point is a function and you can treat it like one talks.tue  
2023/05/25 Johannes Zenn Diffusion Probabilistic Fields talks.tue  
2023/04/27 Tim Xiao

Johannes Zenn		 Trading Information between Latents in Hierarchical Variational Autoencoders

Resampling Gradients Vanish in Differentiable Sequential Monte Carlo Samplers		 talks.tue


talks.tue		 Poster Presentation
2023/04/12 Robert Bamler Finite Volume Neural Network: Modeling Subsurface Contaminant Transport talks.tue  
2023/03/30 Nicolò Zottino Peer-to-Peer Variational Federated Learning Over Arbitrary Graphs talks.tue  
2023/03/16 Johannes Zenn Langevin Diffusion Variational Inference talks.tue  
2023/03/02 Tim Xiao Git Re-Basin: Merging Models modulo Permutation Symmetries talks.tue  
2023/02/23 Tristan Cinquin Understanding Variational Inference in Function-Space talks.tue  
2023/02/09 Alexander Conzelmann Diffusion Probabilistic Modeling for Video Generation N/A