Virtualizing the physical world into virtual models has been a critical technique for robot navigation and planning in the real world. To foster manipulation with articulated objects in everyday life, this work explores building articulation models of indoor scenes through a robot's purposeful interactions in these scenes. Prior work on articulation reasoning primarily focuses on siloed objects of limited categories. To extend to room-scale environments, the robot has to efficiently and effectively explore a large-scale 3D space, locate articulated objects, and infer their articulations. We introduce an interactive perception approach to this task. Our approach, named Ditto in the House, discovers possible articulated objects through affordance prediction, interacts with these objects to produce articulated motions, and infers the articulation properties from the visual observations before and after each interaction. It tightly couples affordance prediction and articulation inference to improve both tasks. We demonstrate the effectiveness of our approach in both simulation and real-world scenes. Code and additional results are available at //ut-austin-rpl.github.io/HouseDitto/
相關內容
Physics-informed neural networks have emerged as a coherent framework for building predictive models that combine statistical patterns with domain knowledge. The underlying notion is to enrich the optimization loss function with known relationships to constrain the space of possible solutions. Hydrodynamic simulations are a core constituent of modern cosmology, while the required computations are both expensive and time-consuming. At the same time, the comparatively fast simulation of dark matter requires fewer resources, which has led to the emergence of machine learning algorithms for baryon inpainting as an active area of research; here, recreating the scatter found in hydrodynamic simulations is an ongoing challenge. This paper presents the first application of physics-informed neural networks to baryon inpainting by combining advances in neural network architectures with physical constraints, injecting theory on baryon conversion efficiency into the model loss function. We also introduce a punitive prediction comparison based on the Kullback-Leibler divergence, which enforces scatter reproduction. By simultaneously extracting the complete set of baryonic properties for the Simba suite of cosmological simulations, our results demonstrate improved accuracy of baryonic predictions based on dark matter halo properties, successful recovery of the fundamental metallicity relation, and retrieve scatter that traces the target simulation's distribution.
Interactive segmentation enables users to segment as needed by providing cues of objects, which introduces human-computer interaction for many fields, such as image editing and medical image analysis. Typically, massive and expansive pixel-level annotations are spent to train deep models by object-oriented interactions with manually labeled object masks. In this work, we reveal that informative interactions can be made by simulation with semantic-consistent yet diverse region exploration in an unsupervised paradigm. Concretely, we introduce a Multi-granularity Interaction Simulation (MIS) approach to open up a promising direction for unsupervised interactive segmentation. Drawing on the high-quality dense features produced by recent self-supervised models, we propose to gradually merge patches or regions with similar features to form more extensive regions and thus, every merged region serves as a semantic-meaningful multi-granularity proposal. By randomly sampling these proposals and simulating possible interactions based on them, we provide meaningful interaction at multiple granularities to teach the model to understand interactions. Our MIS significantly outperforms non-deep learning unsupervised methods and is even comparable with some previous deep-supervised methods without any annotation.
We are interested in pick-and-place style robot manipulation tasks in cluttered and confined 3D workspaces among movable objects that may be rearranged by the robot and may slide, tilt, lean or topple. A recently proposed algorithm, M4M, determines which objects need to be moved and where by solving a Multi-Agent Pathfinding MAPF abstraction of this problem. It then utilises a nonprehensile push planner to compute actions for how the robot might realise these rearrangements and a rigid body physics simulator to check whether the actions satisfy physics constraints encoded in the problem. However, M4M greedily commits to valid pushes found during planning, and does not reason about orderings over pushes if multiple objects need to be rearranged. Furthermore, M4M does not reason about other possible MAPF solutions that lead to different rearrangements and pushes. In this paper, we extend M4M and present Enhanced-M4M (E-M4M) -- a systematic graph search-based solver that searches over orderings of pushes for movable objects that need to be rearranged and different possible rearrangements of the scene. We introduce several algorithmic optimisations to circumvent the increased computational complexity, discuss the space of problems solvable by E-M4M and show that experimentally, both on the real robot and in simulation, it significantly outperforms the original M4M algorithm, as well as other state-of-the-art alternatives when dealing with complex scenes.
Robot manipulation in cluttered scenes often requires contact-rich interactions with objects. It can be more economical to interact via non-prehensile actions, for example, push through other objects to get to the desired grasp pose, instead of deliberate prehensile rearrangement of the scene. For each object in a scene, depending on its properties, the robot may or may not be allowed to make contact with, tilt, or topple it. To ensure that these constraints are satisfied during non-prehensile interactions, a planner can query a physics-based simulator to evaluate the complex multi-body interactions caused by robot actions. Unfortunately, it is infeasible to query the simulator for thousands of actions that need to be evaluated in a typical planning problem as each simulation is time-consuming. In this work, we show that (i) manipulation tasks (specifically pick-and-place style tasks from a tabletop or a refrigerator) can often be solved by restricting robot-object interactions to adaptive motion primitives in a plan, (ii) these actions can be incorporated as subgoals within a multi-heuristic search framework, and (iii) limiting interactions to these actions can help reduce the time spent querying the simulator during planning by up to 40x in comparison to baseline algorithms. Our algorithm is evaluated in simulation and in the real-world on a PR2 robot using PyBullet as our physics-based simulator. Supplementary video: \url{//youtu.be/ABQc7JbeJPM}.
Real-world manipulation problems in heavy clutter require robots to reason about potential contacts with objects in the environment. We focus on pick-and-place style tasks to retrieve a target object from a shelf where some `movable' objects must be rearranged in order to solve the task. In particular, our motivation is to allow the robot to reason over and consider non-prehensile rearrangement actions that lead to complex robot-object and object-object interactions where multiple objects might be moved by the robot simultaneously, and objects might tilt, lean on each other, or topple. To support this, we query a physics-based simulator to forward simulate these interaction dynamics which makes action evaluation during planning computationally very expensive. To make the planner tractable, we establish a connection between the domain of Manipulation Among Movable Objects and Multi-Agent Pathfinding that lets us decompose the problem into two phases our M4M algorithm iterates over. First we solve a multi-agent planning problem that reasons about the configurations of movable objects but does not forward simulate a physics model. Next, an arm motion planning problem is solved that uses a physics-based simulator but does not search over possible configurations of movable objects. We run simulated and real-world experiments with the PR2 robot and compare against relevant baseline algorithms. Our results highlight that M4M generates complex 3D interactions, and solves at least twice as many problems as the baselines with competitive performance.
Scene Graph Generation (SGG) aims to extract <subject, predicate, object> relationships in images for vision understanding. Although recent works have made steady progress on SGG, they still suffer long-tail distribution issues that tail-predicates are more costly to train and hard to distinguish due to a small amount of annotated data compared to frequent predicates. Existing re-balancing strategies try to haddle it via prior rules but are still confined to pre-defined conditions, which are not scalable for various models and datasets. In this paper, we propose a Cross-modal prediCate boosting (CaCao) framework, where a visually-prompted language model is learned to generate diverse fine-grained predicates in a low-resource way. The proposed CaCao can be applied in a plug-and-play fashion and automatically strengthen existing SGG to tackle the long-tailed problem. Based on that, we further introduce a novel Entangled cross-modal prompt approach for open-world predicate scene graph generation (Epic), where models can generalize to unseen predicates in a zero-shot manner. Comprehensive experiments on three benchmark datasets show that CaCao consistently boosts the performance of multiple scene graph generation models in a model-agnostic way. Moreover, our Epic achieves competitive performance on open-world predicate prediction.
Aiming to link natural language descriptions to specific regions in a 3D scene represented as 3D point clouds, 3D visual grounding is a very fundamental task for human-robot interaction. The recognition errors can significantly impact the overall accuracy and then degrade the operation of AI systems. Despite their effectiveness, existing methods suffer from the difficulty of low recognition accuracy in cases of multiple adjacent objects with similar appearances.To address this issue, this work intuitively introduces the human-robot interaction as a cue to facilitate the development of 3D visual grounding. Specifically, a new task termed Embodied Reference Understanding (ERU) is first designed for this concern. Then a new dataset called ScanERU is constructed to evaluate the effectiveness of this idea. Different from existing datasets, our ScanERU is the first to cover semi-synthetic scene integration with textual, real-world visual, and synthetic gestural information. Additionally, this paper formulates a heuristic framework based on attention mechanisms and human body movements to enlighten the research of ERU. Experimental results demonstrate the superiority of the proposed method, especially in the recognition of multiple identical objects. Our codes and dataset are ready to be available publicly.
Dropped into an unknown environment, what should an agent do to quickly learn about the environment and how to accomplish diverse tasks within it? We address this question within the goal-conditioned reinforcement learning paradigm, by identifying how the agent should set its goals at training time to maximize exploration. We propose "Planning Exploratory Goals" (PEG), a method that sets goals for each training episode to directly optimize an intrinsic exploration reward. PEG first chooses goal commands such that the agent's goal-conditioned policy, at its current level of training, will end up in states with high exploration potential. It then launches an exploration policy starting at those promising states. To enable this direct optimization, PEG learns world models and adapts sampling-based planning algorithms to "plan goal commands". In challenging simulated robotics environments including a multi-legged ant robot in a maze, and a robot arm on a cluttered tabletop, PEG exploration enables more efficient and effective training of goal-conditioned policies relative to baselines and ablations. Our ant successfully navigates a long maze, and the robot arm successfully builds a stack of three blocks upon command. Website: //penn-pal-lab.github.io/peg/
Catastrophic forgetting refers to the tendency that a neural network "forgets" the previous learned knowledge upon learning new tasks. Prior methods have been focused on overcoming this problem on convolutional neural networks (CNNs), where the input samples like images lie in a grid domain, but have largely overlooked graph neural networks (GNNs) that handle non-grid data. In this paper, we propose a novel scheme dedicated to overcoming catastrophic forgetting problem and hence strengthen continual learning in GNNs. At the heart of our approach is a generic module, termed as topology-aware weight preserving~(TWP), applicable to arbitrary form of GNNs in a plug-and-play fashion. Unlike the main stream of CNN-based continual learning methods that rely on solely slowing down the updates of parameters important to the downstream task, TWP explicitly explores the local structures of the input graph, and attempts to stabilize the parameters playing pivotal roles in the topological aggregation. We evaluate TWP on different GNN backbones over several datasets, and demonstrate that it yields performances superior to the state of the art. Code is publicly available at \url{//github.com/hhliu79/TWP}.
This paper focuses on two fundamental tasks of graph analysis: community detection and node representation learning, which capture the global and local structures of graphs, respectively. In the current literature, these two tasks are usually independently studied while they are actually highly correlated. We propose a probabilistic generative model called vGraph to learn community membership and node representation collaboratively. Specifically, we assume that each node can be represented as a mixture of communities, and each community is defined as a multinomial distribution over nodes. Both the mixing coefficients and the community distribution are parameterized by the low-dimensional representations of the nodes and communities. We designed an effective variational inference algorithm which regularizes the community membership of neighboring nodes to be similar in the latent space. Experimental results on multiple real-world graphs show that vGraph is very effective in both community detection and node representation learning, outperforming many competitive baselines in both tasks. We show that the framework of vGraph is quite flexible and can be easily extended to detect hierarchical communities.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191