5 papers
Copy First, Translate Later: Interpreting Translation Dynamics in Multilingual Pretraining
Felicia Körner, Maria Matveev, Florian Eichin +3
Large language models exhibit impressive cross-lingual capabilities. However, prior work analyzes this phenomenon through isolated factors and at sparse points during training, lim…
Understanding Multimodal Failure in Action-Chunking Behavioral Cloning
Lorenzo Mazza, Massimiliano Datres, Ariel Rodriguez +3
Behavioral cloning becomes difficult when the same observation admits several valid actions. We study this problem for action-chunking policies and show that different multimodal p…
Complexity Theory meets Ordinary Differential Equations
Adalbert Fono, Noah Wedlich, Holger Boche +1
This contribution investigates the computational complexity of simulating linear ordinary differential equations (ODEs) on digital computers. We provide an exact characterization o…
The Price of Robustness: Stable Classifiers Need Overparameterization
Jonas von Berg, Adalbert Fono, Massimiliano Datres +2
The relationship between overparameterization, stability, and generalization remains incompletely understood in the setting of discontinuous classifiers. We address this gap by est…
Physics-Informed Video Diffusion For Shallow Water Equations
Yang Bai, George Eskandar, Ziyuan Liu +1
Traditional fluid dynamics simulation pipelines combine numerical solvers with rendering, producing highly realistic results but at considerable computational cost. Diffusion-based…