Training general robotic policies from heterogeneous data for different tasks is a significant challenge. Existing robotic datasets vary in different modalities such as color, depth, tactile, and proprioceptive information, and collected in different domains such as simulation, real robots, and human videos. Current methods usually collect and pool all data from one domain to train a single policy to handle such heterogeneity in tasks and domains, which is prohibitively expensive and difficult. In this work, we present a flexible approach, dubbed Policy Composition, to combine information across such diverse modalities and domains for learning scene-level and task-level generalized manipulation skills, by composing different data distributions represented with diffusion models. Our method can use task-level composition for multi-task manipulation and be composed with analytic cost functions to adapt policy behaviors at inference time. We train our method on simulation, human, and real robot data and evaluate in tool-use tasks. The composed policy achieves robust and dexterous performance under varying scenes and tasks and outperforms baselines from a single data source in both simulation and real-world experiments. See //liruiw.github.io/policycomp for more details .
相關內容
The assumption of a static environment is common in many geometric computer vision tasks like SLAM but limits their applicability in highly dynamic scenes. Since these tasks rely on identifying point correspondences between input images within the static part of the environment, we propose a graph neural network-based sparse feature matching network designed to perform robust matching under challenging conditions while excluding keypoints on moving objects. We employ a similar scheme of attentional aggregation over graph edges to enhance keypoint representations as state-of-the-art feature-matching networks but augment the graph with epipolar and temporal information and vastly reduce the number of graph edges. Furthermore, we introduce a self-supervised training scheme to extract pseudo labels for image pairs in dynamic environments from exclusively unprocessed visual-inertial data. A series of experiments show the superior performance of our network as it excludes keypoints on moving objects compared to state-of-the-art feature matching networks while still achieving similar results regarding conventional matching metrics. When integrated into a SLAM system, our network significantly improves performance, especially in highly dynamic scenes.
Data visualization serves as a critical means for presenting data and mining its valuable insights. The task of chart summarization, through natural language processing techniques, facilitates in-depth data analysis of charts. However, there still are notable deficiencies in terms of visual-language matching and reasoning ability for existing approaches. To address these limitations, this study constructs a large-scale dataset of comprehensive chart-caption pairs and fine-tuning instructions on each chart. Thanks to the broad coverage of various topics and visual styles within this dataset, better matching degree can be achieved from the view of training data. Moreover, we propose an innovative chart summarization method, ChartThinker, which synthesizes deep analysis based on chains of thought and strategies of context retrieval, aiming to improve the logical coherence and accuracy of the generated summaries. Built upon the curated datasets, our trained model consistently exhibits superior performance in chart summarization tasks, surpassing 8 state-of-the-art models over 7 evaluation metrics. Our dataset and codes are publicly accessible.
Secure multi-party computation (MPC) techniques can be used to provide data privacy when users query deep neural network (DNN) models hosted on a public cloud. State-of-the-art MPC techniques can be directly leveraged for DNN models that use simple activation functions such as ReLU. However, these techniques are ineffective and/or inefficient for the complex and highly non-linear activation functions used in cutting-edge DNN models. We present Compact, which produces piece-wise polynomial approximations of complex AFs to enable their efficient use with state-of-the-art MPC techniques. Compact neither requires nor imposes any restriction on model training and results in near-identical model accuracy. To achieve this, we design Compact with input density awareness and use an application-specific simulated annealing type optimization to generate computationally more efficient approximations of complex AFs. We extensively evaluate Compact on four different machine-learning tasks with DNN architectures that use popular complex AFs silu, gelu, and mish. Our experimental results show that Compact incurs negligible accuracy loss while being 2x-5x computationally more efficient than state-of-the-art approaches for DNN models with large number of hidden layers. Our work accelerates easy adoption of MPC techniques to provide user data privacy even when the queried DNN models consist of a number of hidden layers and trained over complex AFs.
We introduce LightIt, a method for explicit illumination control for image generation. Recent generative methods lack lighting control, which is crucial to numerous artistic aspects of image generation such as setting the overall mood or cinematic appearance. To overcome these limitations, we propose to condition the generation on shading and normal maps. We model the lighting with single bounce shading, which includes cast shadows. We first train a shading estimation module to generate a dataset of real-world images and shading pairs. Then, we train a control network using the estimated shading and normals as input. Our method demonstrates high-quality image generation and lighting control in numerous scenes. Additionally, we use our generated dataset to train an identity-preserving relighting model, conditioned on an image and a target shading. Our method is the first that enables the generation of images with controllable, consistent lighting and performs on par with specialized relighting state-of-the-art methods.
We introduce a large-scale dataset named MultiGripperGrasp for robotic grasping. Our dataset contains 30.4M grasps from 11 grippers for 345 objects. These grippers range from two-finger grippers to five-finger grippers, including a human hand. All grasps in the dataset are verified in Isaac Sim to classify them as successful and unsuccessful grasps. Additionally, the object fall-off time for each grasp is recorded as a grasp quality measurement. Furthermore, the grippers in our dataset are aligned according to the orientation and position of their palms, allowing us to transfer grasps from one gripper to another. The grasp transfer significantly increases the number of successful grasps for each gripper in the dataset. Our dataset is useful to study generalized grasp planning and grasp transfer across different grippers.
Aerial robots are required to remain operational even in the event of system disturbances, damages, or failures to ensure resilient and robust task completion and safety. One common failure case is propeller damage, which presents a significant challenge in both quantification and compensation. We propose a novel adaptive control scheme capable of detecting and compensating for multi-rotor propeller damages, ensuring safe and robust flight performances. Our control scheme includes an L1 adaptive controller for damage inference and compensation of single or dual propellers, with the capability to seamlessly transition to a fault-tolerant solution in case the damage becomes severe. We experimentally identify the conditions under which the L1 adaptive solution remains preferable over a fault-tolerant alternative. Experimental results validate the proposed approach, demonstrating its effectiveness in running the adaptive strategy in real time on a quadrotor even in case of damage to multiple propellers.
Manual network configuration automation (NCA) tools face significant challenges in versatility and flexibility due to their reliance on extensive domain expertise and manual design, limiting their adaptability to diverse scenarios and complex application needs. This paper introduces PreConfig, an innovative NCA tool that leverages a pretrained language model for automating network configuration tasks. PreConfig is designed to address the complexity and variety of NCA tasks by framing them as text-to-text transformation problems, thus unifying the tasks of configuration generation, translation, and analysis under a single, versatile model. Our approach overcomes existing tools' limitations by utilizing advances in natural language processing to automatically comprehend and generate network configurations without extensive manual re-engineering. We confront the challenges of integrating domain-specific knowledge into pretrained models and the scarcity of supervision data in the network configuration field. Our solution involves constructing a specialized corpus and further pretraining on network configuration data, coupled with a novel data mining technique for generating task supervision data. The proposed model demonstrates robustness in configuration generation, translation, and analysis, outperforming conventional tools in handling complex networking environments. The experimental results validate the effectiveness of PreConfig, establishing a new direction for automating network configuration tasks with pretrained language models.
Information retrieval is a rapidly evolving field. However it still faces significant limitations in the scientific and industrial vast amounts of information, such as semantic divergence and vocabulary gaps in sparse retrieval, low precision and lack of interpretability in semantic search, or hallucination and outdated information in generative models. In this paper, we introduce a two-block approach to tackle these hurdles for long documents. The first block enhances language understanding in sparse retrieval by query expansion to retrieve relevant documents. The second block deepens the result by providing comprehensive and informative answers to the complex question using only the information spread in the long document, enabling bidirectional engagement. At various stages of the pipeline, intermediate results are presented to users to facilitate understanding of the system's reasoning. We believe this bidirectional approach brings significant advancements in terms of transparency, logical thinking, and comprehensive understanding in the field of scientific information retrieval.
Cellular networks are not merely data access networks to the Internet. Their distinct services and ability to form large complex compounds for roaming purposes make them an attractive research target in their own right. Their promise of providing a consistent service with comparable privacy and security across roaming partners falls apart at close inspection. Thus, there is a need for controlled testbeds and measurement tools for cellular access networks doing justice to the technology's unique structure and global scope. Particularly, such measurements suffer from a combinatorial explosion of operators, mobile plans, and services. To cope with these challenges, we built a framework that geographically decouples the SIM from the cellular modem by selectively connecting both remotely. This allows testing any subscriber with any operator at any modem location within minutes without moving parts. The resulting GSM/UMTS/LTE measurement and testbed platform offers a controlled experimentation environment, which is scalable and cost-effective. The platform is extensible and fully open-sourced, allowing other researchers to contribute locations, SIM cards, and measurement scripts. Using the above framework, our international experiments in commercial networks revealed exploitable inconsistencies in traffic metering, leading to multiple phreaking opportunities, i.e., fare-dodging. We also expose problematic IPv6 firewall configurations, hidden SIM card communication to the home network, and fingerprint dial progress tones to track victims across different roaming networks and countries with voice calls.
Foundation models pre-trained on web-scale data are shown to encapsulate extensive world knowledge beneficial for robotic manipulation in the form of task planning. However, the actual physical implementation of these plans often relies on task-specific learning methods, which require significant data collection and struggle with generalizability. In this work, we introduce Robotic Manipulation through Spatial Constraints of Parts (CoPa), a novel framework that leverages the common sense knowledge embedded within foundation models to generate a sequence of 6-DoF end-effector poses for open-world robotic manipulation. Specifically, we decompose the manipulation process into two phases: task-oriented grasping and task-aware motion planning. In the task-oriented grasping phase, we employ foundation vision-language models (VLMs) to select the object's grasping part through a novel coarse-to-fine grounding mechanism. During the task-aware motion planning phase, VLMs are utilized again to identify the spatial geometry constraints of task-relevant object parts, which are then used to derive post-grasp poses. We also demonstrate how CoPa can be seamlessly integrated with existing robotic planning algorithms to accomplish complex, long-horizon tasks. Our comprehensive real-world experiments show that CoPa possesses a fine-grained physical understanding of scenes, capable of handling open-set instructions and objects with minimal prompt engineering and without additional training. Project page: //copa-2024.github.io/
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191