Full Professor, Karlsruhe Institute of Technology
6 papers at NeurIPS 2025
We propose a meta-learning approach to enable Graph Network Simulators (GNSs) for fast adaptation to new physical parameters, leveraging conditional neural processes (CNPs) and neural operators to learn shared latent structures.
We present PointMapPolicy, a multi-modal imitation learning method that conditions diffusion policies on point maps
We iteratively predict target resolutions on intermediate meshes to generate fine-grained adaptive meshes on novel geometries.
We use VLMs to generate high-level hand-object plans for dexterous manipulation.
We combine denosing diffusion probabilistic models and hierarchical graph neural networks to autoregressively simulate physical dynamics on unstructured meshes.
We develop trust region methods for stochastic optimal control to improve sampling from unnormalized densities, transition path sampling, and diffusion model finetuning.
