We introduce LUMOS, a language-conditioned multi-task imitation learning framework for robotics. LUMOS learns skills by practicing them over many long-horizon rollouts in the latent space of a learned world model and transfers these skills zero-shot to a real robot. By learning on-policy in the latent space of the learned world model, our algorithm mitigates policy-induced distribution shift which most offline imitation learning methods suffer from. LUMOS learns from unstructured play data with fewer than 1% hindsight language annotations but is steerable with language commands at test time. We achieve this coherent long-horizon performance by combining latent planning with both image- and language-based hindsight goal relabeling during training, and by optimizing an intrinsic reward defined in the latent space of the world model over multiple time steps, effectively reducing covariate shift. In experiments on the difficult long-horizon CALVIN benchmark, LUMOS outperforms prior learning-based methods with comparable approaches on chained multi-task evaluations. To the best of our knowledge, we are the first to learn a language-conditioned continuous visuomotor control for a real-world robot within an offline world model. Videos, dataset and code are available at http://lumos.cs.uni-freiburg.de.

This paper addresses the challenge of occluded robot grasping, i.e. grasping in situations where the desired grasp poses are kinematically infeasible due to environmental constraints such as surface collisions. Traditional robot manipulation approaches struggle with the complexity of non-prehensile or bimanual strategies commonly used by humans in these circumstances. State-of-the-art reinforcement learning (RL) methods are unsuitable due to the inherent complexity of the task. In contrast, learning from demonstration requires collecting a significant number of expert demonstrations, which is often infeasible. Instead, inspired by human bimanual manipulation strategies, where two hands coordinate to stabilise and reorient objects, we focus on a bimanual robotic setup to tackle this challenge. In particular, we introduce Constraint-based Manipulation for Bimanual Occluded Grasping (COMBO-Grasp), a learning-based approach which leverages two coordinated policies: a constraint policy trained using self-supervised datasets to generate stabilising poses and a grasping policy trained using RL that reorients and grasps the target object. A key contribution lies in value function-guided policy coordination. Specifically, during RL training for the grasping policy, the constraint policy's output is refined through gradients from a jointly trained value function, improving bimanual coordination and task performance. Lastly, COMBO-Grasp employs teacher-student policy distillation to effectively deploy point cloud-based policies in real-world environments. Empirical evaluations demonstrate that COMBO-Grasp significantly improves task success rates compared to competitive baseline approaches, with successful generalisation to unseen objects in both simulated and real-world environments.

--Operators working with robots in safety-critical domains have to make decisions under uncertainty, which remains a challenging problem for a single human operator . An open question is whether two human operators can make better decisions jointly, as compared to a single operator alone. While prior work has shown that two heads are better than one, such studies have been mostly limited to static and passive tasks. We investigate joint decision-making in a dynamic task involving humans teleoperating robots. We conduct a human-subject experiment with N = 100 participants where each participant performed a navigation task with two mobiles robots in simulation. We find that joint decision-making through confidence sharing improves dyad performance beyond the better-performing individual ( p < 0 .0001). Further, we find that the extent of this benefit is regulated both by the skill level of each individual, as well as how well-calibrated their confidence estimates are. Finally, we present findings on characterising the human-human dyad's confidence calibration based on the individuals constituting the dyad. Our findings demonstrate for the first time that two heads are better than one, even on a spatiotemporal task which includes active operator control of robots. I. INTRODUCTION Human operators are increasingly collaborating with robots via teleoperation in domains such as inspection [32, 10, 15, 16, 18, 69], nuclear decommissioning [55, 17], and search and rescue [13, 21, 46, 54]. In these complex environments, operators are often faced with the decision of choosing which robot or robot controller to operate.

We investigate the phenomenon of generalization through the lens of compression. In particular, we study the complexity dynamics of neural networks to explain grokking, where networks suddenly transition from memorizing to generalizing solutions long after over-fitting the training data. To this end we introduce a new measure of intrinsic complexity for neural networks based on the theory of Kolmogorov complexity. Tracking this metric throughout network training, we find a consistent pattern in training dynamics, consisting of a rise and fall in complexity. We demonstrate that this corresponds to memorization followed by generalization. Based on insights from rate--distortion theory and the minimum description length principle, we lay out a principled approach to lossy compression of neural networks, and connect our complexity measure to explicit generalization bounds. Based on a careful analysis of information capacity in neural networks, we propose a new regularization method which encourages networks towards low-rank representations by penalizing their spectral entropy, and find that our regularizer outperforms baselines in total compression of the dataset.

We present a diffusion-based approach to quadrupedal locomotion that simultaneously addresses the limitations of learning and interpolating between multiple skills and of (modes) offline adapting to new locomotion behaviours after training. This is the first framework to apply classifier-free guided diffusion to quadruped locomotion and demonstrate its efficacy by extracting goal-conditioned behaviour from an originally unlabelled dataset. We show that these capabilities are compatible with a multi-skill policy and can be applied with little modification and minimal compute overhead, i.e., running entirely on the robots onboard CPU. We verify the validity of our approach with hardware experiments on the ANYmal quadruped platform.

Causal structures play a central role in world models that flexibly adapt to changes in the environment. While recent works motivate the benefits of discovering local causal graphs for dynamics modelling, in this work we demonstrate that accurately capturing these relationships in complex settings remains challenging for the current state-of-the-art. To remedy this shortcoming, we postulate that sparsity is a critical ingredient for the discovery of such local causal structures. To this end we present the SPARse TrANsformer World model (SPARTAN), a Transformer-based world model that learns local causal structures between entities in a scene. By applying sparsity regularisation on the attention pattern between object-factored tokens, SPARTAN identifies sparse local causal models that accurately predict future object states. Furthermore, we extend our model to capture sparse interventions with unknown targets on the dynamics of the environment. This results in a highly interpretable world model that can efficiently adapt to changes. Empirically, we evaluate SPARTAN against the current state-of-the-art in object-centric world models on observation-based environments and demonstrate that our model can learn accurate local causal graphs and achieve significantly improved few-shot adaptation to changes in the dynamics of the environment as well as robustness against removing irrelevant distractors.

Differentiable simulators continue to push the state of the art across a range of domains including computational physics, robotics, and machine learning. Their main value is the ability to compute gradients of physical processes, which allows differentiable simulators to be readily integrated into commonly employed gradient-based optimization schemes. To achieve this, a number of design decisions need to be considered representing trade-offs in versatility, computational speed, and accuracy of the gradients obtained. This paper presents an in-depth review of the evolving landscape of differentiable physics simulators. We introduce the foundations and core components of differentiable simulators alongside common design choices. This is followed by a practical guide and overview of open-source differentiable simulators that have been used across past research. Finally, we review and contextualize prominent applications of differentiable simulation. By offering a comprehensive review of the current state-of-the-art in differentiable simulation, this work aims to serve as a resource for researchers and practitioners looking to understand and integrate differentiable physics within their research. We conclude by highlighting current limitations as well as providing insights into future directions for the field.

The current state-of-the-art in quadruped locomotion is able to produce robust motion for terrain traversal but requires the segmentation of a desired robot trajectory into a discrete set of locomotion skills such as trot and crawl. In contrast, in this work we demonstrate the feasibility of learning a single, unified representation for quadruped locomotion enabling continuous blending between gait types and characteristics. We present Gaitor, which learns a disentangled representation of locomotion skills, thereby sharing information common to all gait types seen during training. The structure emerging in the learnt representation is interpretable in that it is found to encode phase correlations between the different gait types. These can be leveraged to produce continuous gait transitions. In addition, foot swing characteristics are disentangled and directly addressable. Together with a rudimentary terrain encoding and a learned planner operating in this structured latent representation, Gaitor is able to take motion commands including desired gait type and characteristics from a user while reacting to uneven terrain. We evaluate Gaitor in both simulated and real-world settings on the ANYmal C platform. To the best of our knowledge, this is the first work learning such a unified and interpretable latent representation for multiple gaits, resulting in on-demand continuous blending between different locomotion modes on a real quadruped robot.

We present COmpetitive Mechanisms for Efficient Transfer (COMET), a modular world model which leverages reusable, independent mechanisms across different environments. COMET is trained on multiple environments with varying dynamics via a two-step process: competition and composition. This enables the model to recognise and learn transferable mechanisms. Specifically, in the competition phase, COMET is trained with a winner-takes-all gradient allocation, encouraging the emergence of independent mechanisms. These are then re-used in the composition phase, where COMET learns to re-compose learnt mechanisms in ways that capture the dynamics of intervened environments. In so doing, COMET explicitly reuses prior knowledge, enabling efficient and interpretable adaptation. We evaluate COMET on environments with image-based observations. In contrast to competitive baselines, we demonstrate that COMET captures recognisable mechanisms without supervision. Moreover, we show that COMET is able to adapt to new environments with varying numbers of objects with improved sample efficiency compared to more conventional finetuning approaches.

Mastering dexterous robotic manipulation of deformable objects is vital for overcoming the limitations of parallel grippers in real-world applications. Current trajectory optimisation approaches often struggle to solve such tasks due to the large search space and the limited task information available from a cost function. In this work, we propose D-Cubed, a novel trajectory optimisation method using a latent diffusion model (LDM) trained from a task-agnostic play dataset to solve dexterous deformable object manipulation tasks. D-Cubed learns a skill-latent space that encodes short-horizon actions in the play dataset using a VAE and trains a LDM to compose the skill latents into a skill trajectory, representing a long-horizon action trajectory in the dataset. To optimise a trajectory for a target task, we introduce a novel gradient-free guided sampling method that employs the Cross-Entropy method within the reverse diffusion process. In particular, D-Cubed samples a small number of noisy skill trajectories using the LDM for exploration and evaluates the trajectories in simulation. Then, D-Cubed selects the trajectory with the lowest cost for the subsequent reverse process. This effectively explores promising solution areas and optimises the sampled trajectories towards a target task throughout the reverse diffusion process. Through empirical evaluation on a public benchmark of dexterous deformable object manipulation tasks, we demonstrate that D-Cubed outperforms traditional trajectory optimisation and competitive baseline approaches by a significant margin. We further demonstrate that trajectories found by D-Cubed readily transfer to a real-world LEAP hand on a folding task.