Search

This survey examines the broad suite of methods and models for combining machine learning with physics knowledge for prediction and forecast, with a focus on partial differential equations. These methods have attracted significant interest due to their potential impact on advancing scientific research and industrial practices by improving predictive models with small- or large-scale datasets and expressive predictive models with useful inductive biases. The survey has two parts. The first considers incorporating physics knowledge on an architectural level through objective functions, structured predictive models, and data augmentation. The second considers data as physics knowledge, which motivates looking at multi-task, meta, and contextual learning as an alternative approach to incorporating physics knowledge in a data-driven fashion. Finally, we also provide an industrial perspective on the application of these methods and a survey of the open-source ecosystem for physics-informed machine learning., Comment: 56 pages, 8 figures, 2 tables

Amidst the wide popularity of imitation learning algorithms in robotics, their properties regarding hyperparameter sensitivity, ease of training, data efficiency, and performance have not been well-studied in high-precision industry-inspired environments. In this work, we demonstrate the limitations and benefits of prominent imitation learning approaches and analyze their capabilities regarding these properties. We evaluate each algorithm on a complex bimanual manipulation task involving an over-constrained dynamics system in a setting involving multiple contacts between the manipulated object and the environment. While we find that imitation learning is well suited to solve such complex tasks, not all algorithms are equal in terms of handling environmental and hyperparameter perturbations, training requirements, performance, and ease of use. We investigate the empirical influence of these key characteristics by employing a carefully designed experimental procedure and learning environment. Paper website: https://bimanual-imitation.github.io/

Learning from Demonstrations, the field that proposes to learn robot behavior models from data, is gaining popularity with the emergence of deep generative models. Although the problem has been studied for years under names such as Imitation Learning, Behavioral Cloning, or Inverse Reinforcement Learning, classical methods have relied on models that don't capture complex data distributions well or don't scale well to large numbers of demonstrations. In recent years, the robot learning community has shown increasing interest in using deep generative models to capture the complexity of large datasets. In this survey, we aim to provide a unified and comprehensive review of the last year's progress in the use of deep generative models in robotics. We present the different types of models that the community has explored, such as energy-based models, diffusion models, action value maps, or generative adversarial networks. We also present the different types of applications in which deep generative models have been used, from grasp generation to trajectory generation or cost learning. One of the most important elements of generative models is the generalization out of distributions. In our survey, we review the different decisions the community has made to improve the generalization of the learned models. Finally, we highlight the research challenges and propose a number of future directions for learning deep generative models in robotics., Comment: 20 pages, 11 figures, submitted to TRO

We tackle the recently introduced benchmark for whole-body humanoid control HumanoidBench using MuJoCo MPC. We find that sparse reward functions of HumanoidBench yield undesirable and unrealistic behaviors when optimized; therefore, we propose a set of regularization terms that stabilize the robot behavior across tasks. Current evaluations on a subset of tasks demonstrate that our proposed reward function allows achieving the highest HumanoidBench scores while maintaining realistic posture and smooth control signals. Our code is publicly available and will become a part of MuJoCo MPC, enabling rapid prototyping of robot behaviors., Comment: 3 pages, 3 figures, submitted to IEEE Conference on Robotics and Automation (ICRA@40)

In this work, we introduce PianoMime, a framework for training a piano-playing agent using internet demonstrations. The internet is a promising source of large-scale demonstrations for training our robot agents. In particular, for the case of piano-playing, Youtube is full of videos of professional pianists playing a wide myriad of songs. In our work, we leverage these demonstrations to learn a generalist piano-playing agent capable of playing any arbitrary song. Our framework is divided into three parts: a data preparation phase to extract the informative features from the Youtube videos, a policy learning phase to train song-specific expert policies from the demonstrations and a policy distillation phase to distil the policies into a single generalist agent. We explore different policy designs to represent the agent and evaluate the influence of the amount of training data on the generalization capability of the agent to novel songs not available in the dataset. We show that we are able to learn a policy with up to 56\% F1 score on unseen songs.

Learning a locomotion controller for a musculoskeletal system is challenging due to over-actuation and high-dimensional action space. While many reinforcement learning methods attempt to address this issue, they often struggle to learn human-like gaits because of the complexity involved in engineering an effective reward function. In this paper, we demonstrate that adversarial imitation learning can address this issue by analyzing key problems and providing solutions using both current literature and novel techniques. We validate our methodology by learning walking and running gaits on a simulated humanoid model with 16 degrees of freedom and 92 Muscle-Tendon Units, achieving natural-looking gaits with only a few demonstrations.

Shared dynamics models are important for capturing the complexity and variability inherent in Human-Robot Interaction (HRI). Therefore, learning such shared dynamics models can enhance coordination and adaptability to enable successful reactive interactions with a human partner. In this work, we propose a novel approach for learning a shared latent space representation for HRIs from demonstrations in a Mixture of Experts fashion for reactively generating robot actions from human observations. We train a Variational Autoencoder (VAE) to learn robot motions regularized using an informative latent space prior that captures the multimodality of the human observations via a Mixture Density Network (MDN). We show how our formulation derives from a Gaussian Mixture Regression formulation that is typically used approaches for learning HRI from demonstrations such as using an HMM/GMM for learning a joint distribution over the actions of the human and the robot. We further incorporate an additional regularization to prevent "mode collapse", a common phenomenon when using latent space mixture models with VAEs. We find that our approach of using an informative MDN prior from human observations for a VAE generates more accurate robot motions compared to previous HMM-based or recurrent approaches of learning shared latent representations, which we validate on various HRI datasets involving interactions such as handshakes, fistbumps, waving, and handovers. Further experiments in a real-world human-to-robot handover scenario show the efficacy of our approach for generating successful interactions with four different human interaction partners., Comment: Preprint version of paper accepted at IEEE RAL. Project URL: https://bit.ly/MoVEInt

Prompt learning methods are gaining increasing attention due to their ability to customize large vision-language models to new domains using pre-trained contextual knowledge and minimal training data. However, existing works typically rely on optimizing unified prompt inputs, often struggling with fine-grained classification tasks due to insufficient discriminative attributes. To tackle this, we consider a new framework based on a dual context of both domain-shared and class-specific contexts, where the latter is generated by Large Language Models (LLMs) such as GPTs. Such dual prompt methods enhance the model's feature representation by joining implicit and explicit factors encoded in LLM knowledge. Moreover, we formulate the Unbalanced Optimal Transport (UOT) theory to quantify the relationships between constructed prompts and visual tokens. Through partial matching, UOT can properly align discrete sets of visual tokens and prompt embeddings under different mass distributions, which is particularly valuable for handling irrelevant or noisy elements, ensuring that the preservation of mass does not restrict transport solutions. Furthermore, UOT's characteristics integrate seamlessly with image augmentation, expanding the training sample pool while maintaining a reasonable distance between perturbed images and prompt inputs. Extensive experiments across few-shot classification and adapter settings substantiate the superiority of our model over current state-of-the-art baselines., Comment: Version 1

Planning robot contact often requires reasoning over a horizon to anticipate outcomes, making such planning problems computationally expensive. In this letter, we propose a learning framework for efficient contact planning in real-time subject to uncertain contact dynamics. We implement our approach for the example task of robot air hockey. Based on a learned stochastic model of puck dynamics, we formulate contact planning for shooting actions as a stochastic optimal control problem with a chance constraint on hitting the goal. To achieve online re-planning capabilities, we propose to train an energy-based model to generate optimal shooting plans in real time. The performance of the trained policy is validated %in experiments both in simulation and on a real-robot setup. Furthermore, our approach was tested in a competitive setting as part of the NeurIPS 2023 Robot Air Hockey Challenge., Comment: This work has been submitted to the IEEE Robotics & Automation Letters for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

We present a framework for intuitive robot programming by non-experts, leveraging natural language prompts and contextual information from the Robot Operating System (ROS). Our system integrates large language models (LLMs), enabling non-experts to articulate task requirements to the system through a chat interface. Key features of the framework include: integration of ROS with an AI agent connected to a plethora of open-source and commercial LLMs, automatic extraction of a behavior from the LLM output and execution of ROS actions/services, support for three behavior modes (sequence, behavior tree, state machine), imitation learning for adding new robot actions to the library of possible actions, and LLM reflection via human and environment feedback. Extensive experiments validate the framework, showcasing robustness, scalability, and versatility in diverse scenarios, including long-horizon tasks, tabletop rearrangements, and remote supervisory control. To facilitate the adoption of our framework and support the reproduction of our results, we have made our code open-source. You can access it at: https://github.com/huawei-noah/HEBO/tree/master/ROSLLM., Comment: This document contains 26 pages and 13 figures

Deep Reinforcement Learning (RL) is well known for being highly sensitive to hyperparameters, requiring practitioners substantial efforts to optimize them for the problem at hand. In recent years, the field of automated Reinforcement Learning (AutoRL) has grown in popularity by trying to address this issue. However, these approaches typically hinge on additional samples to select well-performing hyperparameters, hindering sample-efficiency and practicality in RL. Furthermore, most AutoRL methods are heavily based on already existing AutoML methods, which were originally developed neglecting the additional challenges inherent to RL due to its non-stationarities. In this work, we propose a new approach for AutoRL, called Adaptive $Q$-Network (AdaQN), that is tailored to RL to take into account the non-stationarity of the optimization procedure without requiring additional samples. AdaQN learns several $Q$-functions, each one trained with different hyperparameters, which are updated online using the $Q$-function with the smallest approximation error as a shared target. Our selection scheme simultaneously handles different hyperparameters while coping with the non-stationarity induced by the RL optimization procedure and being orthogonal to any critic-based RL algorithm. We demonstrate that AdaQN is theoretically sound and empirically validate it in MuJoCo control problems, showing benefits in sample-efficiency, overall performance, training stability, and robustness to stochasticity., Comment: Preprint

Humans have exceptional tactile sensing capabilities, which they can leverage to solve challenging, partially observable tasks that cannot be solved from visual observation alone. Research in tactile sensing attempts to unlock this new input modality for robots. Lately, these sensors have become cheaper and, thus, widely available. At the same time, the question of how to integrate them into control loops is still an active area of research, with central challenges being partial observability and the contact-rich nature of manipulation tasks. In this study, we propose to use Reinforcement Learning to learn an end-to-end policy, mapping directly from tactile sensor readings to actions. Specifically, we use Dreamer-v3 on a challenging, partially observable robotic insertion task with a Franka Research 3, both in simulation and on a real system. For the real setup, we built a robotic platform capable of resetting itself fully autonomously, allowing for extensive training runs without human supervision. Our preliminary results indicate that Dreamer is capable of utilizing tactile inputs to solve robotic manipulation tasks in simulation and reality. Furthermore, we find that providing the robot with tactile feedback generally improves task performance, though, in our setup, we do not yet include other sensing modalities. In the future, we plan to utilize our platform to evaluate a wide range of other Reinforcement Learning algorithms on tactile tasks.

Telerobotics enables humans to overcome spatial constraints and allows them to physically interact with the environment in remote locations. However, the sensory feedback provided by the system to the operator is often purely visual, limiting the operator's dexterity in manipulation tasks. In this work, we address this issue by equipping the robot's end-effector with high-resolution visuotactile GelSight sensors. Using low-cost MANUS-Gloves, we provide the operator with haptic feedback about forces acting at the points of contact in the form of vibration signals. We propose two different methods for estimating these forces; one based on estimating the movement of markers on the sensor surface and one deep-learning approach. Additionally, we integrate our system into a virtual-reality teleoperation pipeline in which a human operator controls both arms of a Tiago robot while receiving visual and haptic feedback. We believe that integrating haptic feedback is a crucial step for dexterous manipulation in teleoperated robotic systems.

Clustering of motion trajectories is highly relevant for human-robot interactions as it allows the anticipation of human motions, fast reaction to those, as well as the recognition of explicit gestures. Further, it allows automated analysis of recorded motion data. Many clustering algorithms for trajectories build upon distance metrics that are based on pointwise Euclidean distances. However, our work indicates that focusing on salient characteristics is often sufficient. We present a novel distance measure for motion plans consisting of state and control trajectories that is based on a compressed representation built from their main features. This approach allows a flexible choice of feature classes relevant to the respective task. The distance measure is used in agglomerative hierarchical clustering. We compare our method with the widely used dynamic time warping algorithm on test sets of motion plans for the Furuta pendulum and the Manutec robot arm and on real-world data from a human motion dataset. The proposed method demonstrates slight advantages in clustering and strong advantages in runtime, especially for long trajectories., Comment: Published in: 2023 IEEE-RAS 22nd International Conference on Humanoid Robots (Humanoids). Code available at: https://github.com/cztuda/semantic-feature-clustering

Integrating learning-based techniques, especially reinforcement learning, into robotics is promising for solving complex problems in unstructured environments. However, most existing approaches are trained in well-tuned simulators and subsequently deployed on real robots without online fine-tuning. In this setting, the simulation's realism seriously impacts the deployment's success rate. Instead, learning with real-world interaction data offers a promising alternative: not only eliminates the need for a fine-tuned simulator but also applies to a broader range of tasks where accurate modeling is unfeasible. One major problem for on-robot reinforcement learning is ensuring safety, as uncontrolled exploration can cause catastrophic damage to the robot or the environment. Indeed, safety specifications, often represented as constraints, can be complex and non-linear, making safety challenging to guarantee in learning systems. In this paper, we show how we can impose complex safety constraints on learning-based robotics systems in a principled manner, both from theoretical and practical points of view. Our approach is based on the concept of the Constraint Manifold, representing the set of safe robot configurations. Exploiting differential geometry techniques, i.e., the tangent space, we can construct a safe action space, allowing learning agents to sample arbitrary actions while ensuring safety. We demonstrate the method's effectiveness in a real-world Robot Air Hockey task, showing that our method can handle high-dimensional tasks with complex constraints. Videos of the real robot experiments are available on the project website (https://puzeliu.github.io/TRO-ATACOM)., Comment: 19 pages; sumitted to IEEE Transactions on Robotics

This paper explores active sensing strategies that employ vision-based tactile sensors for robotic perception and classification of fabric textures. We formalize the active sampling problem in the context of tactile fabric recognition and provide an implementation of information-theoretic exploration strategies based on minimizing predictive entropy and variance of probabilistic models. Through ablation studies and human experiments, we investigate which components are crucial for quick and reliable texture recognition. Along with the active sampling strategies, we evaluate neural network architectures, representations of uncertainty, influence of data augmentation, and dataset variability. By evaluating our method on a previously published Active Clothing Perception Dataset and on a real robotic system, we establish that the choice of the active exploration strategy has only a minor influence on the recognition accuracy, whereas data augmentation and dropout rate play a significantly larger role. In a comparison study, while humans achieve 66.9% recognition accuracy, our best approach reaches 90.0% in under 5 touches, highlighting that vision-based tactile sensors are highly effective for fabric texture recognition., Comment: 7 pages, 9 figures, accepted at 2024 IEEE International Conference on Robotics and Automation (ICRA)

The vast majority of Reinforcement Learning methods is largely impacted by the computation effort and data requirements needed to obtain effective estimates of action-value functions, which in turn determine the quality of the overall performance and the sample-efficiency of the learning procedure. Typically, action-value functions are estimated through an iterative scheme that alternates the application of an empirical approximation of the Bellman operator and a subsequent projection step onto a considered function space. It has been observed that this scheme can be potentially generalized to carry out multiple iterations of the Bellman operator at once, benefiting the underlying learning algorithm. However, till now, it has been challenging to effectively implement this idea, especially in high-dimensional problems. In this paper, we introduce iterated $Q$-Network (iQN), a novel principled approach that enables multiple consecutive Bellman updates by learning a tailored sequence of action-value functions where each serves as the target for the next. We show that iQN is theoretically grounded and that it can be seamlessly used in value-based and actor-critic methods. We empirically demonstrate the advantages of iQN in Atari $2600$ games and MuJoCo continuous control problems., Comment: Preprint

We consider a sequential decision making task, where the goal is to optimize an unknown function without evaluating parameters that violate an a~priori unknown (safety) constraint. A common approach is to place a Gaussian process prior on the unknown functions and allow evaluations only in regions that are safe with high probability. Most current methods rely on a discretization of the domain and cannot be directly extended to the continuous case. Moreover, the way in which they exploit regularity assumptions about the constraint introduces an additional critical hyperparameter. In this paper, we propose an information-theoretic safe exploration criterion that directly exploits the GP posterior to identify the most informative safe parameters to evaluate. The combination of this exploration criterion with a well known Bayesian optimization acquisition function yields a novel safe Bayesian optimization selection criterion. Our approach is naturally applicable to continuous domains and does not require additional explicit hyperparameters. We theoretically analyze the method and show that we do not violate the safety constraint with high probability and that we learn about the value of the safe optimum up to arbitrary precision. Empirical evaluations demonstrate improved data-efficiency and scalability., Comment: arXiv admin note: text overlap with arXiv:2212.04914

Hidden Markov Models with an underlying Mixture of Gaussian structure have proven effective in learning Human-Robot Interactions from demonstrations for various interactive tasks via Gaussian Mixture Regression. However, a mismatch occurs when segmenting the interaction using only the observed state of the human compared to the joint state of the human and the robot. To enhance this underlying segmentation and subsequently the predictive abilities of such Gaussian Mixture-based approaches, we take a hierarchical approach by learning an additional mixture distribution on the states at the transition boundary. This helps prevent misclassifications that usually occur in such states. We find that our framework improves the performance of the underlying Gaussian Mixture-based approach, which we evaluate on various interactive tasks such as handshaking and fistbumps., Comment: Accepted as a Late Breaking Report in The ACM/IEEE International Conference on Human Robot Interaction (HRI) 2024

Bimanual handovers are crucial for transferring large, deformable or delicate objects. This paper proposes a framework for generating kinematically constrained human-like bimanual robot motions to ensure seamless and natural robot-to-human object handovers. We use a Hidden Semi-Markov Model (HSMM) to reactively generate suitable response trajectories for a robot based on the observed human partner's motion. The trajectories are adapted with task space constraints to ensure accurate handovers. Results from a pilot study show that our approach is perceived as more human--like compared to a baseline Inverse Kinematics approach., Comment: Accepted as a Late Breaking Report in The ACM/IEEE International Conference on Human Robot Interaction (HRI) 2024

A molecule's 2D representation consists of its atoms, their attributes, and the molecule's covalent bonds. A 3D (geometric) representation of a molecule is called a conformer and consists of its atom types and Cartesian coordinates. Every conformer has a potential energy, and the lower this energy, the more likely it occurs in nature. Most existing machine learning methods for molecular property prediction consider either 2D molecular graphs or 3D conformer structure representations in isolation. Inspired by recent work on using ensembles of conformers in conjunction with 2D graph representations, we propose $\mathrm{E}$(3)-invariant molecular conformer aggregation networks. The method integrates a molecule's 2D representation with that of multiple of its conformers. Contrary to prior work, we propose a novel 2D-3D aggregation mechanism based on a differentiable solver for the Fused Gromov-Wasserstein Barycenter problem and the use of an efficient conformer generation method based on distance geometry. We show that the proposed aggregation mechanism is $\mathrm{E}$(3) invariant and propose an efficient GPU implementation. Moreover, we demonstrate that the aggregation mechanism helps to significantly outperform state-of-the-art molecule property prediction methods on established datasets., Comment: Accepted at ICML 2024 (updated version)

Approximate value iteration (AVI) is a family of algorithms for reinforcement learning (RL) that aims to obtain an approximation of the optimal value function. Generally, AVI algorithms implement an iterated procedure where each step consists of (i) an application of the Bellman operator and (ii) a projection step into a considered function space. Notoriously, the Bellman operator leverages transition samples, which strongly determine its behavior, as uninformative samples can result in negligible updates or long detours, whose detrimental effects are further exacerbated by the computationally intensive projection step. To address these issues, we propose a novel alternative approach based on learning an approximate version of the Bellman operator rather than estimating it through samples as in AVI approaches. This way, we are able to (i) generalize across transition samples and (ii) avoid the computationally intensive projection step. For this reason, we call our novel operator projected Bellman operator (PBO). We formulate an optimization problem to learn PBO for generic sequential decision-making problems, and we theoretically analyze its properties in two representative classes of RL problems. Furthermore, we theoretically study our approach under the lens of AVI and devise algorithmic implementations to learn PBO in offline and online settings by leveraging neural network parameterizations. Finally, we empirically showcase the benefits of PBO w.r.t. the regular Bellman operator on several RL problems., Comment: Proceedings of the National Conference on Artificial Intelligence (AAAI-24)

Peer learning is a novel high-level reinforcement learning framework for agents learning in groups. While standard reinforcement learning trains an individual agent in trial-and-error fashion, all on its own, peer learning addresses a related setting in which a group of agents, i.e., peers, learns to master a task simultaneously together from scratch. Peers are allowed to communicate only about their own states and actions recommended by others: "What would you do in my situation?". Our motivation is to study the learning behavior of these agents. We formalize the teacher selection process in the action advice setting as a multi-armed bandit problem and therefore highlight the need for exploration. Eventually, we analyze the learning behavior of the peers and observe their ability to rank the agents' performance within the study group and understand which agents give reliable advice. Further, we compare peer learning with single agent learning and a state-of-the-art action advice baseline. We show that peer learning is able to outperform single-agent learning and the baseline in several challenging discrete and continuous OpenAI Gym domains. Doing so, we also show that within such a framework complex policies from action recommendations beyond discrete action spaces can evolve., Comment: 9 pages, 7 figures, AAAI-24

We consider the problem of quantifying uncertainty over expected cumulative rewards in model-based reinforcement learning. In particular, we focus on characterizing the variance over values induced by a distribution over MDPs. Previous work upper bounds the posterior variance over values by solving a so-called uncertainty Bellman equation (UBE), but the over-approximation may result in inefficient exploration. We propose a new UBE whose solution converges to the true posterior variance over values and leads to lower regret in tabular exploration problems. We identify challenges to apply the UBE theory beyond tabular problems and propose a suitable approximation. Based on this approximation, we introduce a general-purpose policy optimization algorithm, Q-Uncertainty Soft Actor-Critic (QU-SAC), that can be applied for either risk-seeking or risk-averse policy optimization with minimal changes. Experiments in both online and offline RL demonstrate improved performance compared to other uncertainty estimation methods., Comment: arXiv admin note: substantial text overlap with arXiv:2302.12526

Optical tactile sensors have recently become popular. They provide high spatial resolution, but struggle to offer fine temporal resolutions. To overcome this shortcoming, we study the idea of replacing the RGB camera with an event-based camera and introduce a new event-based optical tactile sensor called Evetac. Along with hardware design, we develop touch processing algorithms to process its measurements online at 1000 Hz. We devise an efficient algorithm to track the elastomer's deformation through the imprinted markers despite the sensor's sparse output. Benchmarking experiments demonstrate Evetac's capabilities of sensing vibrations up to 498 Hz, reconstructing shear forces, and significantly reducing data rates compared to RGB optical tactile sensors. Moreover, Evetac's output and the marker tracking provide meaningful features for learning data-driven slip detection and prediction models. The learned models form the basis for a robust and adaptive closed-loop grasp controller capable of handling a wide range of objects. We believe that fast and efficient event-based tactile sensors like Evetac will be essential for bringing human-like manipulation capabilities to robotics. The sensor design is open-sourced at https://sites.google.com/view/evetac ., Comment: Accepted at IEEE Transactions On Robotics. Project Website: https://sites.google.com/view/evetac

Problem gambling is a major public health concern and is associated with profound psychological distress and economic problems. There are numerous gambling communities on the internet where users exchange information about games, gambling tactics, as well as gambling-related problems. Individuals exhibiting higher levels of problem gambling engage more in such communities. Online gambling communities may provide insights into problem-gambling behaviour. Using data scraped from a major German gambling discussion board, we fine-tuned a large language model, specifically a Bidirectional Encoder Representations from Transformers (BERT) model, to predict signs of problem-gambling from forum posts. Training data were generated by manual annotation and by taking into account diagnostic criteria and gambling-related cognitive distortions. Using k-fold cross-validation, our models achieved a precision of 0.95 and F1 score of 0.71, demonstrating that satisfactory classification performance can be achieved by generating high-quality training material through manual annotation based on diagnostic criteria. The current study confirms that a BERT-based model can be reliably used on small data sets and to detect signatures of problem gambling in online communication data. Such computational approaches may have potential for the detection of changes in problem-gambling prevalence among online users.

Multi-Task Reinforcement Learning (MTRL) tackles the long-standing problem of endowing agents with skills that generalize across a variety of problems. To this end, sharing representations plays a fundamental role in capturing both unique and common characteristics of the tasks. Tasks may exhibit similarities in terms of skills, objects, or physical properties while leveraging their representations eases the achievement of a universal policy. Nevertheless, the pursuit of learning a shared set of diverse representations is still an open challenge. In this paper, we introduce a novel approach for representation learning in MTRL that encapsulates common structures among the tasks using orthogonal representations to promote diversity. Our method, named Mixture Of Orthogonal Experts (MOORE), leverages a Gram-Schmidt process to shape a shared subspace of representations generated by a mixture of experts. When task-specific information is provided, MOORE generates relevant representations from this shared subspace. We assess the effectiveness of our approach on two MTRL benchmarks, namely MiniGrid and MetaWorld, showing that MOORE surpasses related baselines and establishes a new state-of-the-art result on MetaWorld., Comment: Accepted at the Twelfth International Conference on Learning Representations (ICLR 2024)

The advent of tactile sensors in robotics has sparked many ideas on how robots can leverage direct contact measurements of their environment interactions to improve manipulation tasks. An important line of research in this regard is that of grasp force control, which aims to manipulate objects safely by limiting the amount of force exerted on the object. While prior works have either hand-modeled their force controllers, employed model-based approaches, or have not shown sim-to-real transfer, we propose a model-free deep reinforcement learning approach trained in simulation and then transferred to the robot without further fine-tuning. We therefore present a simulation environment that produces realistic normal forces, which we use to train continuous force control policies. An evaluation in which we compare against a baseline and perform an ablation study shows that our approach outperforms the hand-modeled baseline and that our proposed inductive bias and domain randomization facilitate sim-to-real transfer. Code, models, and supplementary videos are available on https://sites.google.com/view/rl-force-ctrl

Stateful policies play an important role in reinforcement learning, such as handling partially observable environments, enhancing robustness, or imposing an inductive bias directly into the policy structure. The conventional method for training stateful policies is Backpropagation Through Time (BPTT), which comes with significant drawbacks, such as slow training due to sequential gradient propagation and the occurrence of vanishing or exploding gradients. The gradient is often truncated to address these issues, resulting in a biased policy update. We present a novel approach for training stateful policies by decomposing the latter into a stochastic internal state kernel and a stateless policy, jointly optimized by following the stateful policy gradient. We introduce different versions of the stateful policy gradient theorem, enabling us to easily instantiate stateful variants of popular reinforcement learning and imitation learning algorithms. Furthermore, we provide a theoretical analysis of our new gradient estimator and compare it with BPTT. We evaluate our approach on complex continuous control tasks, e.g., humanoid locomotion, and demonstrate that our gradient estimator scales effectively with task complexity while offering a faster and simpler alternative to BPTT.

Imitation Learning (IL) holds great promise for enabling agile locomotion in embodied agents. However, many existing locomotion benchmarks primarily focus on simplified toy tasks, often failing to capture the complexity of real-world scenarios and steering research toward unrealistic domains. To advance research in IL for locomotion, we present a novel benchmark designed to facilitate rigorous evaluation and comparison of IL algorithms. This benchmark encompasses a diverse set of environments, including quadrupeds, bipeds, and musculoskeletal human models, each accompanied by comprehensive datasets, such as real noisy motion capture data, ground truth expert data, and ground truth sub-optimal data, enabling evaluation across a spectrum of difficulty levels. To increase the robustness of learned agents, we provide an easy interface for dynamics randomization and offer a wide range of partially observable tasks to train agents across different embodiments. Finally, we provide handcrafted metrics for each task and ship our benchmark with state-of-the-art baseline algorithms to ease evaluation and enable fast benchmarking., Comment: https://github.com/robfiras/loco-mujoco

Varying dynamics parameters in simulation is a popular Domain Randomization (DR) approach for overcoming the reality gap in Reinforcement Learning (RL). Nevertheless, DR heavily hinges on the choice of the sampling distribution of the dynamics parameters, since high variability is crucial to regularize the agent's behavior but notoriously leads to overly conservative policies when randomizing excessively. In this paper, we propose a novel approach to address sim-to-real transfer, which automatically shapes dynamics distributions during training in simulation without requiring real-world data. We introduce DOmain RAndomization via Entropy MaximizatiON (DORAEMON), a constrained optimization problem that directly maximizes the entropy of the training distribution while retaining generalization capabilities. In achieving this, DORAEMON gradually increases the diversity of sampled dynamics parameters as long as the probability of success of the current policy is sufficiently high. We empirically validate the consistent benefits of DORAEMON in obtaining highly adaptive and generalizable policies, i.e. solving the task at hand across the widest range of dynamics parameters, as opposed to representative baselines from the DR literature. Notably, we also demonstrate the Sim2Real applicability of DORAEMON through its successful zero-shot transfer in a robotic manipulation setup under unknown real-world parameters., Comment: Published as a conference paper at ICLR 2024. Project website at https://gabrieletiboni.github.io/doraemon/

Robustness against adversarial attacks and distribution shifts is a long-standing goal of Reinforcement Learning (RL). To this end, Robust Adversarial Reinforcement Learning (RARL) trains a protagonist against destabilizing forces exercised by an adversary in a competitive zero-sum Markov game, whose optimal solution, i.e., rational strategy, corresponds to a Nash equilibrium. However, finding Nash equilibria requires facing complex saddle point optimization problems, which can be prohibitive to solve, especially for high-dimensional control. In this paper, we propose a novel approach for adversarial RL based on entropy regularization to ease the complexity of the saddle point optimization problem. We show that the solution of this entropy-regularized problem corresponds to a Quantal Response Equilibrium (QRE), a generalization of Nash equilibria that accounts for bounded rationality, i.e., agents sometimes play random actions instead of optimal ones. Crucially, the connection between the entropy-regularized objective and QRE enables free modulation of the rationality of the agents by simply tuning the temperature coefficient. We leverage this insight to propose our novel algorithm, Quantal Adversarial RL (QARL), which gradually increases the rationality of the adversary in a curriculum fashion until it is fully rational, easing the complexity of the optimization problem while retaining robustness. We provide extensive evidence of QARL outperforming RARL and recent baselines across several MuJoCo locomotion and navigation problems in overall performance and robustness., Comment: Under review

Large, high-capacity models trained on diverse datasets have shown remarkable successes on efficiently tackling downstream applications. In domains from NLP to Computer Vision, this has led to a consolidation of pretrained models, with general pretrained backbones serving as a starting point for many applications. Can such a consolidation happen in robotics? Conventionally, robotic learning methods train a separate model for every application, every robot, and even every environment. Can we instead train generalist X-robot policy that can be adapted efficiently to new robots, tasks, and environments? In this paper, we provide datasets in standardized data formats and models to make it possible to explore this possibility in the context of robotic manipulation, alongside experimental results that provide an example of effective X-robot policies. We assemble a dataset from 22 different robots collected through a collaboration between 21 institutions, demonstrating 527 skills (160266 tasks). We show that a high-capacity model trained on this data, which we call RT-X, exhibits positive transfer and improves the capabilities of multiple robots by leveraging experience from other platforms. More details can be found on the project website https://robotics-transformer-x.github.io., Comment: Project website: https://robotics-transformer-x.github.io

Motion planning is still an open problem for many disciplines, e.g., robotics, autonomous driving, due to their need for high computational resources that hinder real-time, efficient decision-making. A class of methods striving to provide smooth solutions is gradient-based trajectory optimization. However, those methods usually suffer from bad local minima, while for many settings, they may be inapplicable due to the absence of easy-to-access gradients of the optimization objectives. In response to these issues, we introduce Motion Planning via Optimal Transport (MPOT) -- a \textit{gradient-free} method that optimizes a batch of smooth trajectories over highly nonlinear costs, even for high-dimensional tasks, while imposing smoothness through a Gaussian Process dynamics prior via the planning-as-inference perspective. To facilitate batch trajectory optimization, we introduce an original zero-order and highly-parallelizable update rule: the Sinkhorn Step, which uses the regular polytope family for its search directions. Each regular polytope, centered on trajectory waypoints, serves as a local cost-probing neighborhood, acting as a \textit{trust region} where the Sinkhorn Step "transports" local waypoints toward low-cost regions. We theoretically show that Sinkhorn Step guides the optimizing parameters toward local minima regions of non-convex objective functions. We then show the efficiency of MPOT in a range of problems from low-dimensional point-mass navigation to high-dimensional whole-body robot motion planning, evincing its superiority compared to popular motion planners, paving the way for new applications of optimal transport in motion planning., Comment: Published as a conference paper at NeurIPS 2023. Project website: https://sites.google.com/view/sinkhorn-step/

We present improved algorithms with worst-case regret guarantees for the stochastic linear bandit problem. The widely used "optimism in the face of uncertainty" principle reduces a stochastic bandit problem to the construction of a confidence sequence for the unknown reward function. The performance of the resulting bandit algorithm depends on the size of the confidence sequence, with smaller confidence sets yielding better empirical performance and stronger regret guarantees. In this work, we use a novel tail bound for adaptive martingale mixtures to construct confidence sequences which are suitable for stochastic bandits. These confidence sequences allow for efficient action selection via convex programming. We prove that a linear bandit algorithm based on our confidence sequences is guaranteed to achieve competitive worst-case regret. We show that our confidence sequences are tighter than competitors, both empirically and theoretically. Finally, we demonstrate that our tighter confidence sequences give improved performance in several hyperparameter tuning tasks., Comment: Accepted at NeurIPS 2023. 35 pages, 6 figures

Reinforcement Learning (RL) allows learning non-trivial robot control laws purely from data. However, many successful applications of RL have relied on ad-hoc regularizations, such as hand-crafted curricula, to regularize the learning performance. In this paper, we pair a recent algorithm for automatically building curricula with RL on massively parallelized simulations to learn a tracking controller for a spherical pendulum on a robotic arm via RL. Through an improved optimization scheme that better respects the non-Euclidean task structure, we allow the method to reliably generate curricula of trajectories to be tracked, resulting in faster and more robust learning compared to an RL baseline that does not exploit this form of structured learning. The learned policy matches the performance of an optimal control baseline on the real system, demonstrating the potential of curriculum RL to jointly learn state estimation and control for non-linear tracking tasks.

Curriculum reinforcement learning (CRL) allows solving complex tasks by generating a tailored sequence of learning tasks, starting from easy ones and subsequently increasing their difficulty. Although the potential of curricula in RL has been clearly shown in various works, it is less clear how to generate them for a given learning environment, resulting in various methods aiming to automate this task. In this work, we focus on framing curricula as interpolations between task distributions, which has previously been shown to be a viable approach to CRL. Identifying key issues of existing methods, we frame the generation of a curriculum as a constrained optimal transport problem between task distributions. Benchmarks show that this way of curriculum generation can improve upon existing CRL methods, yielding high performance in various tasks with different characteristics.

Many real-world dynamical systems can be described as State-Space Models (SSMs). In this formulation, each observation is emitted by a latent state, which follows first-order Markovian dynamics. A Probabilistic Deep SSM (ProDSSM) generalizes this framework to dynamical systems of unknown parametric form, where the transition and emission models are described by neural networks with uncertain weights. In this work, we propose the first deterministic inference algorithm for models of this type. Our framework allows efficient approximations for training and testing. We demonstrate in our experiments that our new method can be employed for a variety of tasks and enjoys a superior balance between predictive performance and computational budget.

Quantifying uncertainty about a policy's long-term performance is important to solve sequential decision-making tasks. We study the problem from a model-based Bayesian reinforcement learning perspective, where the goal is to learn the posterior distribution over value functions induced by parameter (epistemic) uncertainty of the Markov decision process. Previous work restricts the analysis to a few moments of the distribution over values or imposes a particular distribution shape, e.g., Gaussians. Inspired by distributional reinforcement learning, we introduce a Bellman operator whose fixed-point is the value distribution function. Based on our theory, we propose Epistemic Quantile-Regression (EQR), a model-based algorithm that learns a value distribution function that can be used for policy optimization. Evaluation across several continuous-control tasks shows performance benefits with respect to established model-based and model-free algorithms.

Learning priors on trajectory distributions can help accelerate robot motion planning optimization. Given previously successful plans, learning trajectory generative models as priors for a new planning problem is highly desirable. Prior works propose several ways on utilizing this prior to bootstrapping the motion planning problem. Either sampling the prior for initializations or using the prior distribution in a maximum-a-posterior formulation for trajectory optimization. In this work, we propose learning diffusion models as priors. We then can sample directly from the posterior trajectory distribution conditioned on task goals, by leveraging the inverse denoising process of diffusion models. Furthermore, diffusion has been recently shown to effectively encode data multimodality in high-dimensional settings, which is particularly well-suited for large trajectory dataset. To demonstrate our method efficacy, we compare our proposed method - Motion Planning Diffusion - against several baselines in simulated planar robot and 7-dof robot arm manipulator environments. To assess the generalization capabilities of our method, we test it in environments with previously unseen obstacles. Our experiments show that diffusion models are strong priors to encode high-dimensional trajectory distributions of robot motions.

Bayesian deep learning approaches assume model parameters to be latent random variables and infer posterior distributions to quantify uncertainty, increase safety and trust, and prevent overconfident and unpredictable behavior. However, weight-space priors are model-specific, can be difficult to interpret and are hard to specify. Instead, we apply a Dirichlet prior in predictive space and perform approximate function-space variational inference. To this end, we interpret conventional categorical predictions from stochastic neural network classifiers as samples from an implicit Dirichlet distribution. By adapting the inference, the same function-space prior can be combined with different models without affecting model architecture or size. We illustrate the flexibility and efficacy of such a prior with toy experiments and demonstrate scalability, improved uncertainty quantification and adversarial robustness with large-scale image classification experiments., Comment: Advances in Approximate Bayesian Inference 2023

Graph neural networks are often used to model interacting dynamical systems since they gracefully scale to systems with a varying and high number of agents. While there has been much progress made for deterministic interacting systems, modeling is much more challenging for stochastic systems in which one is interested in obtaining a predictive distribution over future trajectories. Existing methods are either computationally slow since they rely on Monte Carlo sampling or make simplifying assumptions such that the predictive distribution is unimodal. In this work, we present a deep state-space model which employs graph neural networks in order to model the underlying interacting dynamical system. The predictive distribution is multimodal and has the form of a Gaussian mixture model, where the moments of the Gaussian components can be computed via deterministic moment matching rules. Our moment matching scheme can be exploited for sample-free inference, leading to more efficient and stable training compared to Monte Carlo alternatives. Furthermore, we propose structured approximations to the covariance matrices of the Gaussian components in order to scale up to systems with many agents. We benchmark our novel framework on two challenging autonomous driving datasets. Both confirm the benefits of our method compared to state-of-the-art methods. We further demonstrate the usefulness of our individual contributions in a carefully designed ablation study and provide a detailed runtime analysis of our proposed covariance approximations. Finally, we empirically demonstrate the generalization ability of our method by evaluating its performance on unseen scenarios.

We propose a model predictive control approach for autonomous vehicles that exploits learned Gaussian processes for predicting human driving behavior. The proposed approach employs the uncertainty about the GP's prediction to achieve safety. A multi-mode predictive control approach considers the possible intentions of the human drivers. While the intentions are represented by different Gaussian processes, their probabilities foreseen in the observed behaviors are determined by a suitable online classification. Intentions below a certain probability threshold are neglected to improve performance. The proposed multi-mode model predictive control approach with Gaussian process regression support enables repeated feasibility and probabilistic constraint satisfaction with high probability. The approach is underlined in simulation, considering real-world measurements for training the Gaussian processes.

Model-based reinforcement learning is one approach to increase sample efficiency. However, the accuracy of the dynamics model and the resulting compounding error over modelled trajectories are commonly regarded as key limitations. A natural question to ask is: How much more sample efficiency can be gained by improving the learned dynamics models? Our paper empirically answers this question for the class of model-based value expansion methods in continuous control problems. Value expansion methods should benefit from increased model accuracy by enabling longer rollout horizons and better value function approximations. Our empirical study, which leverages oracle dynamics models to avoid compounding model errors, shows that (1) longer horizons increase sample efficiency, but the gain in improvement decreases with each additional expansion step, and (2) the increased model accuracy only marginally increases the sample efficiency compared to learned models with identical horizons. Therefore, longer horizons and increased model accuracy yield diminishing returns in terms of sample efficiency. These improvements in sample efficiency are particularly disappointing when compared to model-free value expansion methods. Even though they introduce no computational overhead, we find their performance to be on-par with model-based value expansion methods. Therefore, we conclude that the limitation of model-based value expansion methods is not the model accuracy of the learned models. While higher model accuracy is beneficial, our experiments show that even a perfect model will not provide an un-rivalled sample efficiency but that the bottleneck lies elsewhere., Comment: Published as a conference paper at ICLR 2023

Recent methods for imitation learning directly learn a $Q$-function using an implicit reward formulation rather than an explicit reward function. However, these methods generally require implicit reward regularization to improve stability and often mistreat absorbing states. Previous works show that a squared norm regularization on the implicit reward function is effective, but do not provide a theoretical analysis of the resulting properties of the algorithms. In this work, we show that using this regularizer under a mixture distribution of the policy and the expert provides a particularly illuminating perspective: the original objective can be understood as squared Bellman error minimization, and the corresponding optimization problem minimizes a bounded $\chi^2$-Divergence between the expert and the mixture distribution. This perspective allows us to address instabilities and properly treat absorbing states. We show that our method, Least Squares Inverse Q-Learning (LS-IQ), outperforms state-of-the-art algorithms, particularly in environments with absorbing states. Finally, we propose to use an inverse dynamics model to learn from observations only. Using this approach, we retain performance in settings where no expert actions are available.

Deployment of Reinforcement Learning (RL) algorithms for robotics applications in the real world requires ensuring the safety of the robot and its environment. Safe Robot RL (SRRL) is a crucial step towards achieving human-robot coexistence. In this paper, we envision a human-centered SRRL framework consisting of three stages: safe exploration, safety value alignment, and safe collaboration. We examine the research gaps in these areas and propose to leverage interactive behaviors for SRRL. Interactive behaviors enable bi-directional information transfer between humans and robots, such as conversational robot ChatGPT. We argue that interactive behaviors need further attention from the SRRL community. We discuss four open challenges related to the robustness, efficiency, transparency, and adaptability of SRRL with interactive behaviors.

We consider the problem of quantifying uncertainty over expected cumulative rewards in model-based reinforcement learning. In particular, we focus on characterizing the variance over values induced by a distribution over MDPs. Previous work upper bounds the posterior variance over values by solving a so-called uncertainty Bellman equation, but the over-approximation may result in inefficient exploration. We propose a new uncertainty Bellman equation whose solution converges to the true posterior variance over values and explicitly characterizes the gap in previous work. Moreover, our uncertainty quantification technique is easily integrated into common exploration strategies and scales naturally beyond the tabular setting by using standard deep reinforcement learning architectures. Experiments in difficult exploration tasks, both in tabular and continuous control settings, show that our sharper uncertainty estimates improve sample-efficiency., Comment: AISTATS 2023

Motion planning is a mature area of research in robotics with many well-established methods based on optimization or sampling the state space, suitable for solving kinematic motion planning. However, when dynamic motions under constraints are needed and computation time is limited, fast kinodynamic planning on the constraint manifold is indispensable. In recent years, learning-based solutions have become alternatives to classical approaches, but they still lack comprehensive handling of complex constraints, such as planning on a lower-dimensional manifold of the task space while considering the robot's dynamics. This paper introduces a novel learning-to-plan framework that exploits the concept of constraint manifold, including dynamics, and neural planning methods. Our approach generates plans satisfying an arbitrary set of constraints and computes them in a short constant time, namely the inference time of a neural network. This allows the robot to plan and replan reactively, making our approach suitable for dynamic environments. We validate our approach on two simulated tasks and in a demanding real-world scenario, where we use a Kuka LBR Iiwa 14 robotic arm to perform the hitting movement in robotic Air Hockey.