Understanding and manipulating deformable objects (e.g., ropes and fabrics) is an essential yet challenging task with broad applications. Difficulties come from complex states and dynamics, diverse configurations and high-dimensional action space of deformable objects. Besides, the manipulation tasks usually require multiple steps to accomplish, and greedy policies may easily lead to local optimal states. Existing studies usually tackle this problem using reinforcement learning or imitating expert demonstrations, with limitations in modeling complex states or requiring hand-crafted expert policies. In this paper, we study deformable object manipulation using dense visual affordance, with generalization towards diverse states, and propose a novel kind of foresightful dense affordance, which avoids local optima by estimating states' values for long-term manipulation. We propose a framework for learning this representation, with novel designs such as multi-stage stable learning and efficient self-supervised data collection without experts. Experiments demonstrate the superiority of our proposed foresightful dense affordance. Project page: //hyperplane-lab.github.io/DeformableAffordance
相關內容
Robots that assist humans will need to interact with articulated objects such as cabinets or microwaves. Early work on creating systems for doing so used proprioceptive sensing to estimate joint mechanisms during contact. However, nowadays, almost all systems use only vision and no longer consider proprioceptive information during contact. We believe that proprioceptive information during contact is a valuable source of information and did not find clear motivation for not using it in the literature. Therefore, in this paper, we create a system that, starting from a given grasp, uses proprioceptive sensing to open cabinets with a position-controlled robot and a parallel gripper. We perform a qualitative evaluation of this system, where we find that slip between the gripper and handle limits the performance. Nonetheless, we find that the system already performs quite well. This poses the question: should we make more use of proprioceptive information during contact in articulated object manipulation systems, or is it not worth the added complexity, and can we manage with vision alone? We do not have an answer to this question, but we hope to spark some discussion on the matter. The codebase and videos of the system are available at //tlpss.github.io/revisiting-proprioception-for-articulated-manipulation/.
Action recognition is an important problem that requires identifying actions in video by learning complex interactions across scene actors and objects. However, modern deep-learning based networks often require significant computation, and may capture scene context using various modalities that further increases compute costs. Efficient methods such as those used for AR/VR often only use human-keypoint information but suffer from a loss of scene context that hurts accuracy. In this paper, we describe an action-localization method, KeyNet, that uses only the keypoint data for tracking and action recognition. Specifically, KeyNet introduces the use of object based keypoint information to capture context in the scene. Our method illustrates how to build a structured intermediate representation that allows modeling higher-order interactions in the scene from object and human keypoints without using any RGB information. We find that KeyNet is able to track and classify human actions at just 5 FPS. More importantly, we demonstrate that object keypoints can be modeled to recover any loss in context from using keypoint information over AVA action and Kinetics datasets.
We present a robot base placement and control method that enables a mobile manipulator to gracefully recover from manipulation failures while performing tasks on-the-move. A mobile manipulator in motion has a limited window to complete a task, unlike when stationary where it can make repeated attempts until successful. Existing approaches to manipulation on-the-move are typically based on open-loop execution of planned trajectories which does not allow the base controller to react to manipulation failures, slowing down or stopping as required. To overcome this limitation, we present a reactive base control method that repeatedly evaluates the best base placement given the robot's current state, the immediate manipulation task, as well as the next part of a multi-step task. The result is a system that retains the reliability of traditional mobile manipulation approaches where the base comes to a stop, but leverages the performance gains available by performing manipulation on-the-move. The controller keeps the base in range of the target for as long as required to recover from manipulation failures while making as much progress as possible toward the next objective. See //benburgesslimerick.github.io/MotM-FailureRecovery for videos of experiments.
Monocular 3D human pose estimation from RGB images has attracted significant attention in recent years. However, recent models depend on supervised training with 3D pose ground truth data or known pose priors for their target domains. 3D pose data is typically collected with motion capture devices, severely limiting their applicability. In this paper, we present a heuristic weakly supervised 3D human pose (HW-HuP) solution to estimate 3D poses in when no ground truth 3D pose data is available. HW-HuP learns partial pose priors from 3D human pose datasets and uses easy-to-access observations from the target domain to estimate 3D human pose and shape in an optimization and regression cycle. We employ depth data for weak supervision during training, but not inference. We show that HW-HuP meaningfully improves upon state-of-the-art models in two practical settings where 3D pose data can hardly be obtained: human poses in bed, and infant poses in the wild. Furthermore, we show that HW-HuP retains comparable performance to cutting-edge models on public benchmarks, even when such models train on 3D pose data.
Safety is often the most important requirement in robotics applications. Nonetheless, control techniques that can provide safety guarantees are still extremely rare for nonlinear systems, such as robot manipulators. A well-known tool to ensure safety is the Viability kernel, which is the largest set of states from which safety can be ensured. Unfortunately, computing such a set for a nonlinear system is extremely challenging in general. Several numerical algorithms for approximating it have been proposed in the literature, but they suffer from the curse of dimensionality. This paper presents a new approach for numerically approximating the viability kernel of robot manipulators. Our approach solves optimal control problems to compute states that are guaranteed to be on the boundary of the set. This allows us to learn directly the set boundary, therefore learning in a smaller dimensional space. Compared to the state of the art on systems up to dimension 6, our algorithm resulted to be more than 2 times as accurate for the same computation time, or 6 times as fast to reach the same accuracy.
Music demixing is the task of separating different tracks from the given single audio signal into components, such as drums, bass, and vocals from the rest of the accompaniment. Separation of sources is useful for a range of areas, including entertainment and hearing aids. In this paper, we introduce two new benchmarks for the sound source separation tasks and compare popular models for sound demixing, as well as their ensembles, on these benchmarks. For the models' assessments, we provide the leaderboard at //mvsep.com/quality_checker/, giving a comparison for a range of models. The new benchmark datasets are available for download. We also develop a novel approach for audio separation, based on the ensembling of different models that are suited best for the particular stem. The proposed solution was evaluated in the context of the Music Demixing Challenge 2023 and achieved top results in different tracks of the challenge. The code and the approach are open-sourced on GitHub.
This paper introduces and explores a new programming paradigm, Model-based Programming, designed to address the challenges inherent in applying deep learning models to real-world applications. Despite recent significant successes of deep learning models across a range of tasks, their deployment in real business scenarios remains fraught with difficulties, such as complex model training, large computational resource requirements, and integration issues with existing programming languages. To ameliorate these challenges, we propose the concept of 'Model-based Programming' and present a novel programming language - M Language, tailored to a prospective model-centered programming paradigm. M Language treats models as basic computational units, enabling developers to concentrate more on crucial tasks such as model loading, fine-tuning, evaluation, and deployment, thereby enhancing the efficiency of creating deep learning applications. We posit that this innovative programming paradigm will stimulate the extensive application and advancement of deep learning technology and provide a robust foundation for a model-driven future.
Transformer-based methods have demonstrated superior performance for monocular 3D object detection recently, which predicts 3D attributes from a single 2D image. Most existing transformer-based methods leverage visual and depth representations to explore valuable query points on objects, and the quality of the learned queries has a great impact on detection accuracy. Unfortunately, existing unsupervised attention mechanisms in transformer are prone to generate low-quality query features due to inaccurate receptive fields, especially on hard objects. To tackle this problem, this paper proposes a novel ``Supervised Scale-constrained Deformable Attention'' (SSDA) for monocular 3D object detection. Specifically, SSDA presets several masks with different scales and utilizes depth and visual features to predict the local feature for each query. Imposing the scale constraint, SSDA could well predict the accurate receptive field of a query to support robust query feature generation. What is more, SSDA is assigned with a Weighted Scale Matching (WSM) loss to supervise scale prediction, which presents more confident results as compared to the unsupervised attention mechanisms. Extensive experiments on ``KITTI'' demonstrate that SSDA significantly improves the detection accuracy especially on moderate and hard objects, yielding SOTA performance as compared to the existing approaches. Code will be publicly available at //github.com/mikasa3lili/SSD-MonoDETR.
Graph Neural Networks (GNNs) draw their strength from explicitly modeling the topological information of structured data. However, existing GNNs suffer from limited capability in capturing the hierarchical graph representation which plays an important role in graph classification. In this paper, we innovatively propose hierarchical graph capsule network (HGCN) that can jointly learn node embeddings and extract graph hierarchies. Specifically, disentangled graph capsules are established by identifying heterogeneous factors underlying each node, such that their instantiation parameters represent different properties of the same entity. To learn the hierarchical representation, HGCN characterizes the part-whole relationship between lower-level capsules (part) and higher-level capsules (whole) by explicitly considering the structure information among the parts. Experimental studies demonstrate the effectiveness of HGCN and the contribution of each component.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191