Multi-task imitation learning (MTIL) has shown significant potential in robotic manipulation by enabling agents to perform various tasks using a single policy. It simplifies the policy deployment and enhances the agent's adaptability across different scenarios. However, key challenges remain, such as maintaining action reliability (e.g., avoiding abnormal action sequences that deviate from nominal task trajectories) and generalizing to unseen tasks with a few expert demonstrations. To address these challenges, we introduce the Foresight-Augmented Manipulation (FoAM) policy, a novel MTIL policy that pioneers the use of multi-modal goal conditions as input and introduces a foresight augmentation in addition to the general action reconstruction. FoAM enables the agent to reason about its actions' visual consequences (foresight) and to be guided by these more expressive representations during task execution. Extensive experiments on over 100 tasks in simulation and real-world settings demonstrate that FoAM significantly enhances MTIL policy performance, outperforming state-of-the-art baselines by up to 41% in success rate. We released our simulation suites that include over 80 challenging tasks across more than 10 scenarios designed for manipulation policy training and evaluation.

3D reconstruction methods such as 3D Gaussian Splatting (3DGS), have achieved significant advancements in recent years. However, the study of articulated objects remains limited due to their geometric and dynamic complexity. We propose Splatter Joint, a novel method that models articulated objects, particularly focusing on joints, to capture both the appearance and the geometric information from a few images taken at a single viewpoint. By integrating joint parameters into the 3DGS rendering process in a differentiable manner, we enable the prediction of joint movements while enhancing the accuracy of object appearance reconstruction. We evaluated Splatter Joint on existing and newly created datasets, demonstrating its effectiveness in modeling object appearance and geometry simultaneously.

In-hand manipulation is challenging in robotics due to the intricate contact dynamics and high degrees of control freedom. Precise manipulation with high accuracy often requires tactile perception, which adds further complexity to the system. Despite the challenges in perception and control, the rolling stick problem is an essential and practical motion primitive with many demanding industrial applications. This work aims to learn the high-resolution tactile dynamics of the rolling stick. Specifically, we try manipulating a small stick using the Allegro hand equipped with the Digit vision-based tactile sensor. The learning framework includes an action filtering module, tactile perception module, and learning with uncertainty module, all designed to operate in low data regimes. With only 2.3% amount of data and 5.7% model complexity of previous similar work, our learned contact dynamics model achieves better grasp stability, sub-millimeter precision, and promising zero-shot generalizability across novel objects. The proposed framework demonstrates the potential for precise in-hand manipulation with tactile feedback on real hardware. The project source code is available at: https://github.com/duyipai/Allegro_Digit. A video presentation is available here.

Manipulating cables is challenging for robots because of the infinite degrees of freedom of the cables and frequent occlusion by the gripper and the environment. These challenges are further complicated by the dexterous nature of the operations required for cable routing and assembly, such as weaving and inserting, hampering common solutions with vision-only sensing. In this paper, we propose to integrate tactile-guided low-level motion control with high-level vision- based task parsing for a challenging task: cable routing and assembly on a reconfigurable task board. Specifically, we build a library of tactile-guided motion primitives using a fingertip GelSight sensor, where each primitive reliably accomplishes an operation such as cable following and weaving. The overall task is inferred via visual perception given a goal configuration image, and then used to generate the primitive sequence. Experiments demonstrate the effectiveness of individual tactile- guided primitives and the integrated end-to-end solution, sig- nificantly outperforming the method without tactile sensing. Our reconfigurable task setup and proposed baselines provide a benchmark for future research in cable manipulation.

Humans are capable of abstracting various tasks as different combinations of multiple attributes. This perspective of compositionality is vital for human rapid learning and adaption since previous experiences from related tasks can be combined to generalize across novel compositional settings. In this work, we aim to achieve zero-shot policy generalization of Reinforcement Learning (RL) agents by leveraging the task compositionality. Our proposed method is a meta-RL algorithm with disentangled task representation, explicitly encoding different aspects of the tasks. Policy generalization is then performed by inferring unseen compositional task representations via the obtained disentanglement without extra exploration. The evaluation is conducted on three simulated tasks and a challenging real-world robotic insertion task. Experimental results demonstrate that our proposed method achieves policy generalization to unseen compositional tasks in a zero-shot manner.

In recent years, impressive results have been achieved in robotic manipulation. While many efforts focus on generating collision-free reference signals, few allow safe contact between the robot bodies and the environment. However, in human's daily manipulation, contact between arms and obstacles is prevalent and even necessary. This paper investigates the benefit of allowing safe contact during robotic manipulation and advocates generating and tracking compliance reference signals in both operational and null spaces. In addition, to optimize the collision-allowed trajectories, we present a hybrid solver that integrates sampling- and gradient-based approaches. We evaluate the proposed method on a goal-reaching task in five simulated and real-world environments with different collisional conditions. We show that allowing safe contact improves goal-reaching efficiency and provides feasible solutions in highly collisional scenarios where collision-free constraints cannot be enforced. Moreover, we demonstrate that planning in null space, in addition to operational space, improves trajectory safety. Further information is available at https://rolandzhu.github.io/ContactReach/.

Manipulation tasks often require a robot to adjust its sensorimotor skills based on the state it finds itself in. Taking peg-in-hole as an example: once the peg is aligned with the hole, the robot should push the peg downwards. While high level execution frameworks such as state machines and behavior trees are commonly used to formalize such decision-making problems, these frameworks require a mechanism to detect the high-level symbolic state. Handcrafting heuristics to identify symbolic states can be brittle, and using data-driven methods can produce noisy predictions, particularly when working with limited datasets, as is common in real-world robotic scenarios. This paper proposes a Bayesian state estimation method to predict symbolic states with predicate classifiers. This method requires little training data and allows fusing noisy observations from multiple sensor modalities. We evaluate our framework on a set of real-world peg-in-hole and connector-socket insertion tasks, demonstrating its ability to classify symbolic states and to generalize to unseen tasks, outperforming baseline methods. We also demonstrate the ability of our method to improve the robustness of manipulation policies on a real robot.

Task-relevant grasping is critical for industrial assembly, where downstream manipulation tasks constrain the set of valid grasps. Learning how to perform this task, however, is challenging, since task-relevant grasp labels are hard to define and annotate. There is also yet no consensus on proper representations for modeling or off-the-shelf tools for performing task-relevant grasps. This work proposes a framework to learn task-relevant grasping for industrial objects without the need of time-consuming real-world data collection or manual annotation. To achieve this, the entire framework is trained solely in simulation, including supervised training with synthetic label generation and self-supervised, hand-object interaction. In the context of this framework, this paper proposes a novel, object-centric canonical representation at the category level, which allows establishing dense correspondence across object instances and transferring task-relevant grasps to novel instances. Extensive experiments on task-relevant grasping of densely-cluttered industrial objects are conducted in both simulation and real-world setups, demonstrating the effectiveness of the proposed framework. Code and data are available at https://sites.google.com/view/catgrasp.

Rewards play a crucial role in reinforcement learning. To arrive at the desired policy, the design of a suitable reward function often requires significant domain expertise as well as trial-and-error. Here, we aim to minimize the effort involved in designing reward functions for contact-rich manipulation tasks. In particular, we provide an approach capable of extracting dense reward functions algorithmically from robots’ high-dimensional observations, such as images and tactile feedback. In contrast to state-of-the-art high-dimensional reward learning methodologies, our approach does not leverage adversarial training, and is thus less prone to the associated training instabilities. Instead, our approach learns rewards by estimating task progress in a self-supervised manner. We demonstrate the effectiveness and efficiency of our approach on two contact-rich manipulation tasks, namely, peg-in-hole and USB insertion. The experimental results indicate that the policies trained with the learned reward function achieves better performance and faster convergence compared to the baselines.

A key challenge towards autonomous multi-part object assembly is robust sensorimotor control under uncertainty. In contrast to previous works that rely on a priori knowledge on whether two parts match, we aim to learn this through physical interaction. We propose a hierarchical approach that enables a robot to autonomously assemble parts while being uncertain about part types and positions. In particular, our probabilistic approach learns a set of differentiable filters that leverage the tactile sensorimotor trace from failed assembly attempts to update its belief about part position and type. This enables a robot to overcome assembly failure. We demonstrate the effectiveness of our approach on a set of object fitting tasks. The experimental results show that the proposed approach achieves higher precision in object position and type estimation, and accomplishes object fitting tasks faster than baselines.

In recent years, many learning based approaches have been studied to realize robotic manipulation and assembly tasks, often including vision and force/tactile feedback. How-ever, it is unclear what the baseline state-of-the-art performance is and what the bottleneck problems are. In this work, we evaluate off-the-shelf (OTS) industrial solutions on a recently introduced benchmark, the National Institute of Standards and Technology (NIST) Assembly Task Board. A set of assembly tasks is introduced and baseline methods are provided to understand their intrinsic difficulty. Multiple sensor-based robotic solutions are then evaluated, including hybrid force/motion control and 2D/3D pattern matching. An end-to-end integrated solution that accomplishes the tasks is also provided.The results and findings throughout the study reveal a few noticeable factors that impede the adoptions of the OTS solutions: dependency on expertise, limited applicability, lack of interoperability, no scene awareness or error recovery mechanisms, and high cost. This paper also provides a first attempt of an objective benchmark performance on the NIST Assembly Task Boards as a reference comparison for future works on this problem.

We introduce the convex factorization machine (CFM), which is a convex variant of the widely used Factorization Machines (FMs). Specifically, we employ a linear+quadratic model and regularize the linear term with the ℓ2-regularizer and the quadratic term with the trace norm regularizer. Then, we formulate the CFM optimization as a semidefinite programming problem and propose an efficient optimization procedure with Hazan's algorithm. A key advantage of CFM over existing FMs is that it can find a globally optimal solution, while FMs may get a poor locally optimal solution since the objective function of FMs is non-convex. In addition, the proposed algorithm is simple yet effective and can be implemented easily. Finally, CFM is a general factorization method and can also be used for other factorization problems, including multi-view matrix factorization and tensor completion problems, in various domains including toxicogenomics and bioinformatics. Through synthetic and traditionally used movielens datasets, we first show that the proposed CFM achieves results competitive to FMs. We then show in a toxicogenomics prediction task that CFM predicts the toxic outcomes of a collection of drugs better than a state-of-the-art tensor factorization method.

We present a probabilistic framework for learning with heterogeneous multiview data where some views are given as ordinal, binary, or real-valued feature matrices, and some views as similarity matrices. Our framework has the following distinguishing aspects: (i) a unified latent factor model for integrating information from diverse feature (ordinal, binary, real) and similarity based views, and predicting the missing data in each view, leveraging view correlations; (ii) seamless adaptation to binary/multiclass classification where data consists of multiple feature and/or similarity-based views; and (iii) an efficient, variational inference algorithm which is especially flexible in modeling the views with ordinal-valued data (by learning the cutpoints for the ordinal data), and extends naturally to streaming data settings. Our framework subsumes methods such as multiview learning and multiple kernel learning as special cases. We demonstrate the effectiveness of our framework on several real-world and benchmarks datasets.

Point process data are commonly observed in fields like healthcare and social science. Designing predictive models for such event streams is an under-explored problem, due to often scarce training data. In this work we propose a multitask point process model, leveraging information from all tasks via a hierarchical Gaussian process (GP). Nonparametric learning functions implemented by a GP, which map from past events to future rates, allow analysis of flexible arrival patterns. To facilitate efficient inference, we propose a sparse construction for this hierarchical model, and derive a variational Bayes method for learning and inference. Experimental results are shown on both synthetic data and an application on real electronic health records.

The analysis and characterization of correlated point process data has wide applications, ranging from biomedical research to network analysis. In this work, we model such data as generated by a latent collection of continuous-time binary semi-Markov processes, corresponding to external events appearing and disappearing. A continuous-time modeling framework is more appropriate for multichannel point process data than a binning approach requiring time discretization, and we show connections between our model and recent ideas from the discrete-time literature. We describe an efficient MCMC algorithm for posterior inference, and apply our ideas to both synthetic data and a real-world biometrics application.