Marvin Alles

Nutan Chen, Jianxiang Feng, Marvin Alles et al.

In recent years Flow Matching has become a prominent method for generative modeling robot motion generation. In its generic form Flow Matching is an ODE-based neural sampler that is trained by regressing empirical flow fields associated with motion samples as data. However, in robot motion generation we often have additional constraints that might not be present in the collected data. The majority of current approaches train the flow on the available data and use inference-time guidance to enforce task-specific constraints. To address this mismatch, we propose \textbf{ConFlow}, a constraint-guided flow matching framework that incorporates constraint information directly into the training objective via differentiable barrier or cost functions. To address design specifications such as smoothness and boundary conditions, we propose replacing the standard Gaussian source distribution used in flow matching training with a conditional Gaussian Process. Our approach also uses infeasible demonstrations as negative supervision, improving constraint satisfaction without requiring additional expert data. Experiments on a two-robot navigation task demonstrate that ConFlow achieves lower collision rates and higher trajectory quality than standard flow matching baselines, with or without inference-time guidance. These results validate training-time constraint integration as an effective approach to closing the training--inference gap in generative motion models.

Marvin Alles, Xingyuan Zhang, Patrick van der Smagt et al.

Inspired by how humans combine direct interaction with action-free experience (e.g., videos), we study world models that learn from heterogeneous data. Standard world models typically rely on action-conditioned trajectories, which limits effectiveness when action labels are scarce. We introduce a family of latent-action world models that jointly use action-conditioned and action-free data by learning a shared latent action representation. This latent space aligns observed control signals with actions inferred from passive observations, enabling a single dynamics model to train on large-scale unlabeled trajectories while requiring only a small set of action-labeled ones. We use the latent-action world model to learn a latent-action policy through offline reinforcement learning (RL), thereby bridging two traditionally separate domains: offline RL, which typically relies on action-conditioned data, and action-free training, which is rarely used with subsequent RL. On the DeepMind Control Suite, our approach achieves strong performance while using about an order of magnitude fewer action-labeled samples than purely action-conditioned baselines. These results show that latent actions enable training on both passive and interactive data, which makes world models learn more efficiently.

Marvin Alles, Philip Becker-Ehmck, Patrick van der Smagt et al.

In offline reinforcement learning, a policy is learned using a static dataset in the absence of costly feedback from the environment. In contrast to the online setting, only using static datasets poses additional challenges, such as policies generating out-of-distribution samples. Model-based offline reinforcement learning methods try to overcome these by learning a model of the underlying dynamics of the environment and using it to guide policy search. It is beneficial but, with limited datasets, errors in the model and the issue of value overestimation among out-of-distribution states can worsen performance. Current model-based methods apply some notion of conservatism to the Bellman update, often implemented using uncertainty estimation derived from model ensembles. In this paper, we propose Constrained Latent Action Policies (C-LAP) which learns a generative model of the joint distribution of observations and actions. We cast policy learning as a constrained objective to always stay within the support of the latent action distribution, and use the generative capabilities of the model to impose an implicit constraint on the generated actions. Thereby eliminating the need to use additional uncertainty penalties on the Bellman update and significantly decreasing the number of gradient steps required to learn a policy. We empirically evaluate C-LAP on the D4RL and V-D4RL benchmark, and show that C-LAP is competitive to state-of-the-art methods, especially outperforming on datasets with visual observations.

Marvin Alles, Nutan Chen, Patrick van der Smagt et al.

The use of guidance to steer sampling toward desired outcomes has been widely explored within diffusion models, especially in applications such as image and trajectory generation. However, incorporating guidance during training remains relatively underexplored. In this work, we introduce energy-guided flow matching, a novel approach that enhances the training of flow models and eliminates the need for guidance at inference time. We learn a conditional velocity field corresponding to the flow policy by approximating an energy-guided probability path as a Gaussian path. Learning guided trajectories is appealing for tasks where the target distribution is defined by a combination of data and an energy function, as in reinforcement learning. Diffusion-based policies have recently attracted attention for their expressive power and ability to capture multi-modal action distributions. Typically, these policies are optimized using weighted objectives or by back-propagating gradients through actions sampled by the policy. As an alternative, we propose FlowQ, an offline reinforcement learning algorithm based on energy-guided flow matching. Our method achieves competitive performance while the policy training time is constant in the number of flow sampling steps.

Marvin Alles, Elie Aljalbout

Robotic manipulators are widely used in modern manufacturing processes. However, their deployment in unstructured environments remains an open problem. To deal with the variety, complexity, and uncertainty of real-world manipulation tasks, it is essential to develop a flexible framework with reduced assumptions on the environment characteristics. In recent years, reinforcement learning (RL) has shown great results for single-arm robotic manipulation. However, research focusing on dual-arm manipulation is still rare. From a classical control perspective, solving such tasks often involves complex modeling of interactions between two manipulators and the objects encountered in the tasks, as well as the two robots coupling at a control level. Instead, in this work, we explore the applicability of model-free RL to dual-arm assembly. As we aim to contribute towards an approach that is not limited to dual-arm assembly, but dual-arm manipulation in general, we keep modeling efforts at a minimum. Hence, to avoid modeling the interaction between the two robots and the used assembly tools, we present a modular approach with two decentralized single-arm controllers which are coupled using a single centralized learned policy. We reduce modeling effort to a minimum by using sparse rewards only. Our architecture enables successful assembly and simple transfer from simulation to the real world. We demonstrate the effectiveness of the framework on dual-arm peg-in-hole and analyze sample efficiency and success rates for different action spaces. Moreover, we compare results on different clearances and showcase disturbance recovery and robustness, when dealing with position uncertainties. Finally we zero-shot transfer policies trained in simulation to the real world and evaluate their performance.