Gerhard Neumann

TROLL: Trust Regions Improve Reinforcement Learning for Large Language Models

Learning Boltzmann Generators via Constrained Mass Transport

Chunking the Critic: A Transformer-based Soft Actor-Critic with N-Step Returns

Improving Long-Range Interactions in Graph Neural Simulators via Hamiltonian Dynamics

Fourier Features Let Agents Learn High Precision Policies with Imitation Learning

PAWS: Preference Learning with Advantage-Weighted Segments

Meta-Learning Contextual Time Series Forecasting with Neural Processes

Trust Region Constrained Measure Transport in Path Space for Stochastic Optimal Control and Inference

Geometry-aware RL for Manipulation of Varying Shapes and Deformable Objects

Diffusion-Based Hierarchical Graph Neural Networks for Simulating Nonlinear Solid Mechanics

TOP-ERL: Transformer-based Off-Policy Episodic Reinforcement Learning

AMBER: Adaptive Mesh Generation by Iterative Mesh Resolution Prediction

DIME: Diffusion-Based Maximum Entropy Reinforcement Learning

MaNGO — Adaptable Graph Network Simulators via Meta-Learning

PointMapPolicy: Structured Point Cloud Processing for Multi-Modal Imitation Learning

Underdamped Diffusion Bridges with Applications to Sampling

Efficient Off-Policy Learning for High-Dimensional Action Spaces

End-to-end Learning of Gaussian Mixture Priors for Diffusion Sampler

Sequential Controlled Langevin Diffusions

Scaffolding Dexterous Manipulation with Vision-Language Models

Latent Task-Specific Graph Network Simulators

Perlin Noise for Exploration in Reinforcement Learning