检索结果-内蒙古大学图书馆

From Alexnet to Transformers: Measuring the Non-linearity of Deep Neural Networks with Affine Optimal Transport

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Bouniot, Quentin Redko, Ievgen Mallasto, Anton Laclau, Charlotte Arndt, Karol Struckmeier, Oliver Heinonen, Markus Kyrki, Ville Kaski, Samuel LTCI Télécom Paris Institut Polytechnique de Paris France Noah’s Ark Lab Paris France Smartly.io Finland Aalto University Finland Intelligent Robotics Group Finland Department of Computer Science University of Manchester United Kingdom

In the last decade, we have witnessed the introduction of several novel deep neural network (DNN) architectures exhibiting ever-increasing performance across diverse tasks. Explaining the upward trend of their performance, however, remains difficult as different DNN architectures of comparable depth and width – common factors associated with their expressive power – may exhibit a drastically different performance even when trained on the same dataset. In this paper, we introduce the concept of the non-linearity signature of DNN, the first theoretically sound solution for approximately measuring the non-linearity of deep neural networks. Built upon a score derived from closed-form optimal transport mappings, this signature provides a better understanding of the inner workings of a wide range of DNN architectures and learning paradigms, with a particular emphasis on the computer vision task. We provide extensive experimental results that highlight the practical usefulness of the proposed non-linearity signature and its potential for long-reaching implications. The code for our work is available at https://***/qbouniot/AffScoreDeep. Copyright © 2023, The Authors. All rights reserved.

关键词： Deep neural networks

Industrial machine vision: lessons and challenges

学校读者我要写书评

暂无评论

Optics News 1987年第2期13卷 12-17页

作者： Steven W. Holland Robert B. Tilove STEVEN W. HOLLAND is an assistant department head in the Computer Science Dept. of General Motors Research Laboratories (GMR). He is responsible for research projects involving computer vision robotics mobile robotics process planning knowledge systems natural language advanced programming environments parallel processing and statistics. ROBERT B. TILOVE is senior staff research scientist and project leader of the machine perception project at GMR. His research interests include geometric modeling model-based programming and control of intelligent automation systems.

No abstract available.

关键词： Integral imaging Machine vision Measurement and metrology Optical systems Sensors Statistics

Reactive High-level Behavior Synthesis for an Atlas Humanoid Robot

学校读者我要写书评

暂无评论

Reactive High-level Behavior Synthesis for an Atlas Humanoid...

IEEE International Conference on robotics and Automation

作者： Spyros Maniatopoulos Philipp Schillinger Vitchyr Pong David C. Conner Hadas Kress-Gazit Verifiable Robotics Research Group Sibley School of Mechanical and Aerospace Engineering Cornell University Ithaca NY 14853 USA Robert Bosch GmbH Corporate Research Department for Cognitive Systems 70442 Stuttgart Germany Capable Humanitarian Robotics & Intelligent Systems Lab Department of Physics Computer Science and Engineering Christopher Newport University Newport News VA 23606 USA

ISBN: (纸本)9781467380270

In this work, we take a step towards bridging the gap between the theory of formal synthesis and its application to real-world, complex, robotic systems. In particular, we present an end-to-end approach for the automatic generation of code that implements high-level robot behaviors in a verifiably correct manner, including reaction to the possible failures of low-level actions. We start with a description of the system defined a priori. Thus, a non-expert user need only specify a high-level task. We automatically construct a formal specification, in a fragment of Linear Temporal Logic (LTL), that encodes the system's capabilities and constraints, the task, and the desired reaction to low-level failures. We then synthesize a reactive mission plan that is guaranteed to satisfy the formal specification, i.e., achieve the task's goals or correctly react to failures. Lastly, we automatically generate a state machine that instantiates the synthesized symbolic plan in software. We showcase our approach using Team ViGIR's software and Atlas humanoid robot and present lab experiments, thus demonstrating the application of formal synthesis techniques to complex robotic systems. The proposed approach has been implemented and open-sourced as a collection of Robot Operating System (ROS) packages, which are adaptable to other systems.

关键词： Atlas Humanoid Robot formal synthesis ROS

Constrained Generative Sampling of 6-DoF Grasps

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Lundell, Jens Verdoja, Francesco Le, Tran Nguyen Mousavian, Arsalan Fox, Dieter Kyrki, Ville Intelligent Robotics Group Department of Electrical Engineering and Automation School of Electrical Engineering Aalto University Finland KTH Royal Institute of Technology Sweden NVIDIA Corporation United States Paul G. Allen School of Computer Science & Engineering University of Washington Seattle United States

Most state-of-the-art data-driven grasp sampling methods propose stable and collision-free grasps uniformly on the target object. For bin-picking, executing any of those reachable grasps is sufficient. However, for completing specific tasks, such as squeezing out liquid from a bottle, we want the grasp to be on a specific part of the object’s body while avoiding other locations, such as the cap. This work presents a generative grasp sampling network, VCGS, capable of constrained 6-Degrees of Freedom (DoF) grasp sampling. In addition, we also curate a new dataset designed to train and evaluate methods for constrained grasping. The new dataset, called CONG, consists of over 14 million training samples of synthetically rendered point clouds and grasps at random target areas on 2889 objects. VCGS is benchmarked against GraspNet, a state-of-the-art unconstrained grasp sampler, in simulation and on a real robot. The results demonstrate that VCGS achieves a 10–15% higher grasp success rate than the baseline while being 2–3 times as sample efficient. Supplementary material is available on our project website. Copyright © 2023, The Authors. All rights reserved.

关键词： Bottles

Integrated Bi-Manual Motion Generation and Control shaped for Probabilistic Movement Primitives

学校读者我要写书评

暂无评论

Integrated Bi-Manual Motion Generation and Control shaped fo...

IEEE-RAS International Conference on Humanoid Robots

作者： Jonathan Vorndamme João Carvalho Riddhiman Laha Dorothea Koert Luis Figueredo Jan Peters Sami Haddadin Department of Computer Engineering Technische Universität München (TUM) TUM School of Computation Information and Technology (CIT) Munich Institute of Robotics and Machine Intelligence (MIRMI) Germany Intelligent Autonomous Systems group Technische Universitat Darmstadt (TUD) Germany Center for Cognitive Science TUD Germany

ISBN: (纸本)9798350309805

This work introduces a novel cooperative control framework that allows for real-time reactiveness and adaptation whilst satisfying implicit constraints stemming from proba-bilistic/stochastic trajectories. Stemming from task-oriented sampling and/or task-oriented demonstrations, e.g., learning based on motion primitives, such trajectories carry additional information often neglected during real-time control deployment. In particular, methods such as probabilistic movement primitives offer the advantage to capture the inherent stochasticity in human demonstrations - which in turn reflects human's understanding about task-variability and adaption possibilities. This information, however, is often poorly exploited and, mostly, used during offline trajectory planning stage. Our work instead introduces a novel real-time motion-generation strategy that explicitly exploits such information to improve trajectories according to changes in the environmental condition and robot task-space topology. The proposed solution is particularly well-suited for bi-manual and coordinated systems where the increased kinematic complexity, tightly-coupled constraints and reduced workspace have detrimental effects on the manipula-bility, joint-limits, and are even capable of causing unstable behavior and task-failure. Our methodology addresses these challenges, and improves performance and task-execution by taking the confidence range region explicitly into account whilst maneuvering towards better configurations. Furthermore, it can directly cope with different closed-chain kinematics and task-space topologies, resulting for instance from different grasps. Experimental evaluations on a bi-manual Franka panda robot show that the method can run in the inner control loop of the robot and enables successful execution of highly constrained tasks.

关键词： Trajectory planning Robot kinematics Humanoid robots Kinematics Probabilistic logic Real-time systems Trajectory

Scientific and Technological Challenges in RoboCup

学校读者我要写书评

暂无评论

Control, robotics, and Autonomous Systems 1000年第1期3卷 441-471页

作者： Minoru Asada Oskar von Stryk 1Symbiotic Intelligent Systems Research Center Institute for Open and Transdisciplinary Research Initiatives Osaka University Suita Osaka 565-0871 Japan email: asada@otri.osaka-u.ac.jp 2Simulation Systems Optimization and Robotics Group Department of Computer Science Technische Universität Darmstadt D-64289 Darmstadt Germany

Since its inception in 1997, RoboCup has developed into a truly unique and long-standing research community advancing robotics and artificial intelligence through various challenges, benchmarks, and test fields. The main purposes of this article are to evaluate the research and development achievements so far and to identify new challenges and related new research issues. Unlike other robot competitions and research conferences, RoboCup eliminates the boundaries between pure research activities and the development of full system designs with hardware and software implementations at a site open to the public. It also creates specific scientific and technological research and development challenges to be addressed. In this article, we provide an overview of RoboCup, including its league structure and related research issues. We also review recent studies across several research categories to show how participants (called RoboCuppers) address the research and development challenges before, during, and after the annual competitions. Among the diversity of research issues, we highlight two unique aspects of the challenges: the platform design of the robots and the game evaluations. Both of these aspects contribute to solving the research and development challenges of RoboCup and verifying the results from a common perspective (i.e., a more objective view). Finally, we provide concluding remarks and discuss future research directions.

关键词：

Line Drawing Guided Progressive Inpainting of Mural Damage

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Li, Luxi Zou, Qin Zhang, Fan Yu, Hongkai Chen, Long Song, Chengfang Huang, Xianfeng Wang, Xiaoguang Li, Qingquan Department of Computer Science Technology United International College of Beijing Normal University HongKong Baptist University Zhuhai China Machine Vision and Robotics Laboratory School of Computer Science Wuhan University Wuhan China State Key Laboratory of Surveying Mapping and Remote Sensing Information Engineering Wuhan University Wuhan China Department of Electrical Engineering and Computer Science Cleveland State University OH United States Institute of Automation Chinese Academy of Sciences Beijing China Cultural Heritage Intelligent Computing Laboratory Wuhan University Wuhan China Guangming Laboratory Shenzhen University Shenzhen China

Mural image inpainting is far less explored compared to its natural image counterpart and remains largely unsolved. Most existing image-inpainting methods tend to take the target image as the only input and directly repair the damage to generate a visually plausible result. These methods obtain high performance in restoration or completion of some pre-defined objects, e.g., human face, fabric texture, and printed texts, etc., however, are not suitable for repairing murals with varying subjects and large damaged areas. Moreover, due to discrete colors in paints, mural inpainting may suffer from apparent color bias. To this end, in this paper, we propose a line drawing guided progressive mural inpainting method. It divides the inpainting process into two steps: structure reconstruction and color correction, implemented by a structure reconstruction network (SRN) and a color correction network (CCN), respectively. In structure reconstruction, SRN utilizes the line drawing as an assistant to achieve large-scale content authenticity and structural stability. In color correction, CCN operates a local color adjustment for missing pixels which reduces the negative effects of color bias and edge jumping. The proposed approach is evaluated against the current state-of-the-art image inpainting methods. Qualitative and quantitative results demonstrate the superiority of the proposed method in mural image inpainting. The codes and data are available at https://***/qinnzou/mural-image-inpainting . © 2022, CC BY-NC-ND.

关键词： Restoration

Dynamic process initial conditions in repetitive processes. Controllability and stability analysis

学校读者我要写书评

暂无评论

Dynamic process initial conditions in repetitive processes. ...

Information, Decision and Control (IDC)

作者： K. Galkowski E. Rogers A. Gramacki J. Gramacki D.H. Owens Institute of Robotics and Software Engineering Technical University of Zielona Góra Zielona Gora Poland Department of Electronics and Computer Science Image Speech and Intelligent Systems Research Group University of Southampton Southampton UK Institute of Computer Engineering and Electronics Technical University of Zielona Góra Zielona Gora Poland School of Engineering Centre for Systems and Control Engineering University of Exeter Exeter UK

Repetitive, or multipass, processes are a class of 2D systems of both practical and algorithmic/theoretical interest whose dynamics cannot be analysed or controlled using standard (1D) systems theory. Recently it has been shown that the modelling of the boundary conditions, also known as the process initial conditions, is a crucial feature in the analysis and control of these processes. This paper presents some further results on the effects of so-called 'dynamic' process initial conditions on the controllability and stability properties of discrete linear repetitive processes. Previous work has shown that these dynamic process initial conditions alone can destroy the stability properties of these processes. Hence their effects must be 'adequately' accounted for the process modelling stage in order to ensure that subsequent analysis does not lead to incorrect results/conclusions. The main results developed in this paper can be summarised as follows. (i) Computationally efficient stability tests which can, in effect, be applied using standard, or 1D, linear systems tests. (ii) Characterisation of so-called pass controllability in the form of matrix rank based conditions. (iii) Conditions under which the dynamic process initial conditions can be selected to ensure stability and pass controllability.

关键词： Stability Controllability System testing Control system analysis Algorithm design and analysis Control systems Boundary conditions Process control Standards development Linear systems

CodeEnhance: A Codebook-Driven Approach for Low-Light Image Enhancement

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wu, Xu Hou, XianXu Lai, Zhihui Zhou, Jie Zhang, Ya-Nan Pedrycz, Witold Shen, Linlin The Computer Vision Institute College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen518060 China Guangdong Key Laboratory of Intelligent Information Processing Shenzhen University Shenzhen518060 China School of AI and Advanced Computing Xi’an Jiaotong-Liverpool University China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University SZU Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society Guangdong Shenzhen518060 China The Department of Electrical & Computer Engineering University of Alberta University of Alberta Canada

Low-light image enhancement (LLIE) aims to improve low-illumination images. However, existing methods face two challenges: (1) uncertainty in restoration from diverse brightness degradations;(2) loss of texture and color information caused by noise suppression and light enhancement. In this paper, we propose a novel enhancement approach, CodeEnhance, by leveraging quantized priors and image refinement to address these challenges. In particular, we reframe LLIE as learning an image-to-code mapping from low-light images to discrete codebook, which has been learned from high-quality images. To enhance this process, a Semantic Embedding Module (SEM) is introduced to integrate semantic information with low-level features, and a Codebook Shift (CS) mechanism, designed to adapt the pre-learned codebook to better suit the distinct characteristics of our low-light dataset. Additionally, we present an Interactive Feature Transformation (IFT) module to refine texture and color information during image reconstruction, allowing for interactive enhancement based on user preferences. Extensive experiments on both real-world and synthetic benchmarks demonstrate that the incorporation of prior knowledge and controllable information transfer significantly enhances LLIE performance in terms of quality and fidelity. The proposed CodeEnhance exhibits superior robustness to various degradations, including uneven illumination, noise, and color distortion. Copyright © 2024, The Authors. All rights reserved.

关键词： Image enhancement