activity
20242026
collaborators

5 papers

cs.CL2026

Shieldstral

Antonia Calvi, Avinash Sooriyarachchi, Giada Pistilli +274

We introduce Shieldstral, a 3B-parameter policy-adaptive multimodal safety classifier that matches or outperforms models nearly 7 its size on text safety benchmarks and set…

cs.LG2026

Tadpole: Autoencoders as Foundation Models for 3D PDEs with Online Learning

Qiang Liu, Felix Koehler, Benjamin Holzschuh +1

We introduce Tadpole, a novel foundation model for three-dimensional partial differential equations (PDEs) that addresses key challenges in transferability, scalability to high dim…

cs.LG2025

P3D: Scalable Neural Surrogates for High-Resolution 3D Physics Simulations with Global Context

Benjamin Holzschuh, Georg Kohl, Florian Redinger +1

We present a scalable framework for learning deterministic and probabilistic neural surrogates for high-resolution 3D physics simulations. We introduce a hybrid CNN-Transformer bac…

cs.LG2025

PDE-Transformer: Efficient and Versatile Transformers for Physics Simulations

Benjamin Holzschuh, Qiang Liu, Georg Kohl +1

We introduce PDE-Transformer, an improved transformer-based architecture for surrogate modeling of physics simulations on regular grids. We combine recent architectural improvement…

cs.LG2024

Flow Matching for Posterior Inference with Simulator Feedback

Benjamin Holzschuh, Nils Thuerey

Flow-based generative modeling is a powerful tool for solving inverse problems in physical sciences that can be used for sampling and likelihood evaluation with much lower inferenc…