Diffusion Sequence Models for Generative In-Context Meta-Learning of Robot Dynamics

arXiv:2604.13366v1 Announce Type: new Abstract: Accurate modeling of robot dynamics is essential for model-based control, yet remains challenging under distributional shifts and real-time constraints. In this work, we formulate system identification as an in-context meta-learning problem and compare deterministic and generative sequence models for forward dynamics prediction. We take a Transformer-based meta-model, as a strong deterministic baseline, and introduce to this setting two complementary diffusion-based approaches: (i) inpainting diffusion (Diffuser), which learns the joint input-observation distribution, and (ii) conditioned diffusion models (CNN and Transformer), which generate future observations conditioned on control inputs. Through large-scale randomized simulations, we analyze performance across in-distribution and out-of-distribution regimes, as well as computational trade-offs relevant for control. We show that diffusion models significantly improve robustness under distribution shift, with inpainting diffusion achieving the best performance in our experiments. Finally, we demonstrate that warm-started sampling enables diffusion models to operate within real-time constraints, making them viable for control applications. These results highlight generative meta-models as a promising direction for robust system identification in robotics.

Related

• Binomial Gradient-Based Meta-Learning for Enhanced Meta-Gradient Estimation
• Learning Discrete Diffusion of Graphs via Free-Energy Gradient Flows
• Signal-Aware Conditional Diffusion Surrogates for Transonic Wing Pressure Prediction
• A Bayesian Perspective on the Role of Epistemic Uncertainty for Delayed Generalization in In-Context Learning
• Adjoint Matching through the Lens of the Stochastic Maximum Principle in Optimal Control

Source: arXiv cs.LG | 2026-04-16