AIConfPaper
← Search

Nils Thuerey

36 accepted papers

2026

On the Generalization in Topology Optimization via Sensitivity-Conditioned Bernoulli Flow Matching

Mohammad Rashed, Duarte Filipe Valoroso Madeira, Babak Gholami, Caglar Guerbuez, Yunjia Yang, Nils Thuerey
ICML 2026poster

Surrogate models for topology optimization (TO) exhibit highly variable out-of-distribution (OOD) generalization under distribution shifts such as changing loads or boundary conditions, yet the source of this variability remains unclear. We hypothesize that OOD performance is governed by how much in…

Cited by 0Source ↗Scholar ↗
2026

P3D: Highly Scalable 3D Neural Surrogates for Physics Simulations with Global Context

Benjamin Holzschuh, Georg Kohl, Florian Redinger, Nils Thuerey
ICLR 2026poster

We present a scalable framework for learning deterministic and probabilistic neural surrogates for high-resolution 3D physics simulations. We introduce P3D, a hybrid CNN-Transformer backbone architecture targeted for 3D physics simulations, which significantly outperforms existing architectures in t…

Cited by 0Source ↗code ↗Scholar ↗
2026

Physics vs Distributions: Pareto Optimal Flow Matching with Physics Constraints

Giacomo Baldan, Qiang Liu, Alberto Guardone, Nils Thuerey
ICLR 2026poster

Physics-constrained generative modeling aims to produce high-dimensional samples that are both physically consistent and distributionally accurate, a task that remains challenging due to often conflicting optimization objectives. Recent advances in flow matching and diffusion models have enabled eff…

Cited by 0Source ↗code ↗Scholar ↗
2026

Plug-and-Play Benchmarking of Reinforcement Learning Algorithms for Large-Scale Flow Control

Jannis Becktepe, Aleksandra Franz, Nils Thuerey, Sebastian Peitz
ICML 2026poster

Reinforcement learning (RL) has shown promising results in active flow control (AFC), yet progress in the field remains difficult to assess as existing studies rely on heterogeneous observation and actuation schemes, numerical setups, and evaluation protocols. Current AFC benchmarks attempt to addre…

Cited by 0Source ↗Scholar ↗
2025

ConFIG: Towards Conflict-free Training of Physics Informed Neural Networks

Qiang Liu, Mengyu Chu, Nils Thuerey
ICLR 2025spotlight

The loss functions of many learning problems contain multiple additive terms that can disagree and yield conflicting update directions. For Physics-Informed Neural Networks (PINNs), loss terms on initial/boundary conditions and physics equations are particularly interesting as they are well-establis…

Cited by 10Source ↗PDF ↗code ↗Scholar ↗
2025

INC: An Indirect Neural Corrector for Auto-Regressive Hybrid PDE Solvers

Hao Wei, Aleksandra Franz, Björn Malte List, Nils Thuerey
NeurIPS 2025poster

When simulating partial differential equations, hybrid solvers combine coarse numerical solvers with learned correctors. They promise accelerated simulations while adhering to physical constraints. However, as shown in our theoretical framework, directly applying learned corrections to solver output…

Cited by 0Source ↗Scholar ↗
2025

Improved Sampling Of Diffusion Models In Fluid Dynamics With Tweedie's Formula

Youssef Shehata, Benjamin Holzschuh, Nils Thuerey
ICLR 2025poster

State-of-the-art Denoising Diffusion Probabilistic Models (DDPMs) rely on an expensive sampling process with a large Number of Function Evaluations (NFEs) to provide high-fidelity predictions. This computational bottleneck renders diffusion models less appealing as surrogates for the spatio-temporal…

Cited by 1Source ↗PDF ↗Scholar ↗
2025

Learning Distributions of Complex Fluid Simulations with Diffusion Graph Networks

Mario Lino Valencia, Tobias Pfaff, Nils Thuerey
ICLR 2025oral

Physical systems with complex unsteady dynamics, such as fluid flows, are often poorly represented by a single mean solution. For many practical applications, it is crucial to access the full distribution of possible states, from which relevant statistics (e.g., RMS and two-point correlations) can b…

Cited by 2Source ↗PDF ↗Scholar ↗
2025

Neural Emulator Superiority: When Machine Learning for PDEs Surpasses its Training Data

Felix Koehler, Nils Thuerey
NeurIPS 2025poster

Neural operators or emulators for PDEs trained on data from numerical solvers are conventionally assumed to be limited by their training data's fidelity. We challenge this assumption by identifying "emulator superiority," where neural networks trained purely on low-fidelity solver data can achieve h…

Cited by 0Source ↗code ↗Scholar ↗
2025

PDE-Transformer: Efficient and Versatile Transformers for Physics Simulations

Benjamin Holzschuh, Qiang Liu, Georg Kohl, Nils Thuerey
ICML 2025poster

We introduce PDE-Transformer, an improved transformer-based architecture for surrogate modeling of physics simulations on regular grids. We combine recent architectural improvements of diffusion transformers with adjustments specific for large-scale simulations to yield a more scalable and versatile…

Cited by 0Source ↗PDF ↗code ↗Scholar ↗
2025

Temporal Difference Learning: Why It Can Be Fast and How It Will Be Faster

Patrick Schnell, Luca Guastoni, Nils Thuerey
ICLR 2025poster

Temporal difference (TD) learning represents a fascinating paradox: It is the prime example of a divergent algorithm that has not vanished after its instability was proven. On the contrary, TD continues to thrive in reinforcement learning (RL), suggesting that it provides significant compensatory be…

Cited by 0Source ↗PDF ↗Scholar ↗
2024

APEBench: A Benchmark for Autoregressive Neural Emulators of PDEs

Felix Koehler, Simon Niedermayr, rüdiger westermann, Nils Thuerey
NeurIPS 2024poster

We introduce the **A**utoregressive **P**DE **E**mulator Benchmark (APEBench), a comprehensive benchmark suite to evaluate autoregressive neural emulators for solving partial differential equations. APEBench is based on JAX and provides a seamlessly integrated differentiable simulation framework em…

Cited by 1Source ↗PDF ↗code ↗Scholar ↗
2023

Learning Similarity Metrics for Volumetric Simulations with Multiscale CNNs

Georg Kohl, Li-Wei Chen, Nils Thuerey
AAAI 2023technical

Simulations that produce three-dimensional data are ubiquitous in science, ranging from fluid flows to plasma physics. We propose a similarity model based on entropy, which allows for the creation of physically meaningful ground truth distances for the similarity assessment of scalar and vectorial d…

Cited by 2Source ↗PDF ↗code ↗Scholar ↗
2023

Learning to Estimate Single-View Volumetric Flow Motions without 3D Supervision

Aleksandra Franz, Barbara Solenthaler, Nils Thuerey
ICLR 2023poster

We address the challenging problem of jointly inferring the 3D flow and volumetric densities moving in a fluid from a monocular input video with a deep neural network. Despite the complexity of this task, we show that it is possible to train the corresponding networks without requiring any 3D ground…

Cited by 5Source ↗PDF ↗code ↗Scholar ↗
2022

A Two-stage Learning Architecture that Generates High-Quality Grasps for a Multi-Fingered Hand

Dominik Winkelbauer, Berthold Bäuml, Matthias Humt, Nils Thuerey, Rudolph Triebel
IROS 2022poster

We investigate the problem of planning stable grasps for object manipulations using an 18-DOF robotic hand with four fingers. The main challenge here is the high-dimensional search space, and we address this problem using a novel two-stage learning process. In the first stage, we train an autoregres…

Cited by 12Source ↗Scholar ↗
2022

Guaranteed Conservation of Momentum for Learning Particle-based Fluid Dynamics

Lukas Prantl, Benjamin Ummenhofer, Vladlen Koltun, Nils Thuerey
NeurIPS 2022accept

We present a novel method for guaranteeing linear momentum in learned physics simulations. Unlike existing methods, we enforce conservation of momentum with a hard constraint, which we realize via antisymmetrical continuous convolutional layers. We combine these strict constraints with a hierarchica…

Cited by 32Source ↗PDF ↗code ↗Scholar ↗
2022

ULNeF: Untangled Layered Neural Fields for Mix-and-Match Virtual Try-On

Igor Santesteban, Miguel A. Otaduy, Nils Thuerey, Dan Casas
NeurIPS 2022accept

Recent advances in neural models have shown great results for virtual try-on (VTO) problems, where a 3D representation of a garment is deformed to fit a target body shape. However, current solutions are limited to a single garment layer, and cannot address the combinatorial complexity of mixing diff…

Cited by 39Source ↗PDF ↗Scholar ↗
2021

Global Transport for Fluid Reconstruction With Learned Self-Supervision

Erik Franz, Barbara Solenthaler, Nils Thuerey
CVPR 2021poster

We propose a novel method to reconstruct volumetric flows from sparse views via a global transport formulation. Instead of obtaining the space-time function of the observations, we reconstruct its motion based on a single initial state. In addition we introduce a learned self-supervision that constr…

Cited by 26PDF ↗code ↗Scholar ↗
2021

Self-Supervised Collision Handling via Generative 3D Garment Models for Virtual Try-On

Igor Santesteban, Nils Thuerey, Miguel A. Otaduy, Dan Casas
CVPR 2021poster

We propose a new generative model for 3D garment deformations that enables us to learn, for first time, a data-driven method for virtual try-on that effectively addresses garment-body collisions. In contrast to existing methods that require an undesirable postprocessing step to fix garment-body inte…

Cited by 109PDF ↗code ↗Scholar ↗
2020

Correspondence-Free Material Reconstruction using Sparse Surface Constraints

Sebastian Weiss, Robert Maier, Daniel Cremers, Rudiger Westermann, Nils Thuerey
CVPR 2020poster

We present a method to infer physical material parameters, and even external boundaries, from the scanned motion of a homogeneous deformable object via the solution of an inverse problem. Parameters are estimated from real-world data sources such as sparse observations from a Kinect sensor without c…

Cited by 18PDF ↗Scholar ↗
2020

Lagrangian Fluid Simulation with Continuous Convolutions

Benjamin Ummenhofer, Lukas Prantl, Nils Thuerey, Vladlen Koltun
ICLR 2020poster

We present an approach to Lagrangian fluid simulation with a new type of convolutional network. Our networks process sets of moving particles, which describe fluids in space and time. Unlike previous approaches, we do not build an explicit graph structure to connect the particles but use spatial con…

Cited by 229Source ↗Scholar ↗
2020

Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers

Kiwon Um, Robert Brand, Yun (Raymond) Fei, Philipp Holl, Nils Thuerey
NeurIPS 2020poster

Finding accurate solutions to partial differential equations (PDEs) is a crucial task in all scientific and engineering disciplines. It has recently been shown that machine learning methods can improve the solution accuracy by correcting for effects not captured by the discretized PDE. We target the…

Cited by 327Source ↗PDF ↗code ↗Scholar ↗
2020

Tranquil Clouds: Neural Networks for Learning Temporally Coherent Features in Point Clouds

Lukas Prantl, Nuttapong Chentanez, Stefan Jeschke, Nils Thuerey
ICLR 2020spotlight

Point clouds, as a form of Lagrangian representation, allow for powerful and flexible applications in a large number of computational disciplines. We propose a novel deep-learning method to learn stable and temporally coherent feature spaces for points clouds that change over time. We identify a set…

Cited by 19Source ↗Scholar ↗