检索结果-内蒙古大学图书馆

arXiv 2017年

作者： Rajasekaran, Siddharthan Zhang, Jinwei Fu, Jie Department of Electrical and Computer Engineering Robotics Engineering Program Worcester Polytechnic Institute

Inverse Reinforcement Learning (IRL) is the task of learning a single reward function given a Markov Decision Process (MDP) without defining the reward function, and a set of demonstrations generated by humans/experts. However, in practice, it may be unreasonable to assume that human behaviors can be explained by one reward function since they may be inherently inconsistent. Also, demonstrations may be collected from various users and aggregated to infer and predict users behaviors. In this paper, we introduce the Non-parametric Behavior Clustering IRL algorithm to simultaneously cluster demonstrations and learn multiple reward functions from demonstrations that may be generated from more than one behaviors. Our method is iterative: It alternates between clustering demonstrations into different behavior clusters and inverse learning the reward functions until convergence. It is built upon the Expectation-Maximization formulation and non-parametric clustering in the IRL setting. Further, to improve the computation efficiency, we remove the need of completely solving multiple IRL problems for multiple clusters during the iteration steps and introduce a resampling technique to avoid generating too many unlikely clusters. We demonstrate the convergence and efficiency of the proposed method through learning multiple driver behaviors from demonstrations generated from a grid-world environment and continuous trajectories collected from autonomous robot cars using the Gazebo robot simulator. Copyright © 2017, The Authors. All rights reserved.

关键词： Demonstrations

来源：评论

学校读者我要写书评

暂无评论

DispSegNet: Leveraging semantics for end-to-end learning of disparity estimation from stereo imagery

arXiv

引用

arXiv 2018年

作者： Zhang, Junming Skinner, Katherine A. Vasudevan, Ram Johnson-Roberson, Matthew Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109 United States Robotics Program University of Michigan Ann ArborMI48109 United States Department of Mechanical Engineering University of Michigan Ann ArborMI48109 United States Department of Naval Architecture and Marine Engineering University of Michigan Ann ArborMI48109 United States

Recent work has shown that convolutional neural networks (CNNs) can be applied successfully in disparity estimation, but these methods still suffer from errors in regions of low-texture, occlusions and reflections. Concurrently, deep learning for semantic segmentation has shown great progress in recent years. In this paper, we design a CNN architecture that combines these two tasks to improve the quality and accuracy of disparity estimation with the help of semantic segmentation. Specifically, we propose a network structure in which these two tasks are highly coupled. One key novelty of this approach is the two-stage refinement process. Initial disparity estimates are refined with an embedding learned from the semantic segmentation branch of the network. The proposed model is trained using an unsupervised approach, in which images from one half of the stereo pair are warped and compared against images from the other camera. Another key advantage of the proposed approach is that a single network is capable of outputting disparity estimates and semantic labels. These outputs are of great use in autonomous vehicle operation;with real-time constraints being key, such performance improvements increase the viability of driving applications. Experiments on KITTI and Cityscapes datasets show that our model can achieve state-of-the-art results and that leveraging embedding learned from semantic segmentation improves the performance of disparity estimation. Copyright © 2018, The Authors. All rights reserved.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

CWave: High-performance single-source any-angle path planning on a grid

CWave: High-performance single-source any-angle path plannin...

引用

IEEE International Conference on robotics and Automation (ICRA)

作者： Dmitry A. Sinyukov Taşkin Padir Robotics Engineering Program Worcester Polytechnic Institute Worcester MA USA Department of Electrical and Computer Engineering Northeastern University Boston MA

Path planning on a 2D-grid is a well-studied problem in robotics. It usually involves searching for a shortest path between two vertices on a grid. Single-source path planning is a modified problem which asks to find distances from a given point to all other points on the map. A high-performance algorithm for single-source any-angle path planning on a grid that we named CWave is proposed in this work. “Any-angle” attribute of a path planning algorithm implies that such algorithm can find paths which may include any angle segments, as opposed to standard A* on an 8-connected graph, the path can turn with 45°-increments only. The key idea of the presented algorithm is that it does not represent the grid as a graph and uses discrete geometric primitives to define the wave front. In its purest form, CWave requires for computation only integer arithmetics and multiplication by two, but can accumulate the distance error at turning points. A modified version of CWave with minimal usage of floating-point calculations is also developed. It allows to eliminate any accumulative errors which is proven mathematically and experimentally on several maps. The performance of the algorithm on three maps is demonstrated to be significantly faster than that of Theta*, Lazy Theta* and Field A* adapted for single-source planning. The limitations of the current implementations of the algorithm as well as potential improvements are discussed.

关键词： Path planning Algorithm design and analysis Two dimensional displays Navigation Planning Mobile robots

来源：评论

学校读者我要写书评

暂无评论

NANOROBOTS: Intracellular manipulation and measurement with multipole magnetic tweezers

引用

Science robotics 2019年第28期4卷

作者： Wang, X. Ho, C. Tsatskis, Y. Law, J. Zhang, Z. Zhu, M. Dai, C. Wang, F. Tan, M. Hopyan, S. McNeill, H. Sun, Y. Department of Mechanical and Industrial Engineering University of Toronto TorontoONM5S 3G8 Canada Institute of Biomaterials and Biomedical Engineering TorontoONM5S 3G9 Canada Lunenfeld-Tanenbaum Research Institute Mt. Sinai Hospital TorontoONM5G 1X5 Canada Program in Developmental and Stem Cell Biology Hospital for Sick Children TorontoONM5G 1X8 Canada HRG Central Institute of Robotics HIT Robot Group Harbin Heilongjiang150001 China Institute of Automation Chinese Academy of Sciences Beijing100190 China Division of Orthopaedic Surgery Hospital for Sick Children and University of Toronto TorontoONM5G 1X8 Canada Department of Developmental Biology Washington University School of Medicine St. LouisMO63108 United States Department of Electrical and Computer Engineering University of Toronto TorontoONM5S 3G4 Canada

The capability to directly interrogate intracellular structures inside a single cell for measurement and manipulation is important for understanding subcellular and suborganelle activities, diagnosing diseases, and developing new therapeutic approaches. Compared with measurements of single cells, physical measurement and manipulation of subcellular structures and organelles remain underexplored. To improve intracellular physical measurement and manipulation,we have developed amultipole magnetic tweezers system for micromanipulation involving submicrometer position control and piconewton force control of a submicrometermagnetic bead inside a single cell for measurement in different locations (spatial) and different time points (temporal). The beadwas three-dimensionally positioned in the cell using a generalized predictive controller that addresses the control challenge caused by the low bandwidth of visual feedback from high-resolution confocal imaging. The average positioning error was quantified to be 0.4 mm, slightly larger than the Brownianmotion-imposed constraint (0.31 mm). The system is also capable of applying a force up to 60 pN with a resolution of 4 pN for a period of time longer than 30 min. The measurement results revealed that significantly higher stiffness exists in the nucleus' major axis than in the minor axis. This stiffness polarity is likely attributed to the aligned actin filament. We also showed that the nucleus stiffens upon the application of an intracellularly applied force, which can be attributed to the response of structural protein lamin A/C and the intracellular stress fiber actin filaments. © 2019 The Authors.

关键词： Cells

来源：评论

学校读者我要写书评

暂无评论

Driving in the Matrix: Can virtual worlds replace human-generated annotations for real world tasks?

Driving in the Matrix: Can virtual worlds replace human-gene...

引用

2017 IEEE International Conference on robotics and Automation, ICRA 2017

作者： Johnson-Roberson, Matthew Barto, Charles Mehta, Rounak Sridhar, Sharath Nittur Rosaen, Karl Vasudevan, Ram Department of Naval Architecture and Marine Engineering University of Michigan Ann ArborMI48109 United States Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109 United States Robotics Program University of Michigan Ann ArborMI48109 United States Mechanical Engineering University of Michigan Ann ArborMI48109 United States

ISBN: (纸本)9781509046331

Deep learning has rapidly transformed the state of the art algorithms used to address a variety of problems in computer vision and robotics. These breakthroughs have relied upon massive amounts of human annotated training data. This time consuming process has begun impeding the progress of these deep learning efforts. This paper describes a method to incorporate photo-realistic computer images from a simulation engine to rapidly generate annotated data that can be used for the training of machine learning algorithms. We demonstrate that a state of the art architecture, which is trained only using these synthetic annotations, performs better than the identical architecture trained on human annotated real-world data, when tested on the KITTI data set for vehicle detection. By training machine learning algorithms on a rich virtual world, real objects in real scenes can be learned and classified using synthetic data. This approach offers the possibility of accelerating deep learning's application to sensor-based classification problems like those that appear in self-driving cars. The source code and data to train and validate the networks described in this paper are made available for researchers. © 2017 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

A study on robust control method of underwater vehicle using SMCSPO

A study on robust control method of underwater vehicle using...

引用

Asian Control Conference

作者： Hyun Hee Kim Jong Seob Won Hyeon Jin Cho Jun Ho Hwang Min Cheol Lee Interdisciplinary Program in Robotics Pusan National University Geumjeong-gu Busan Korea Department of Mechanical and Automotive Engineering Jeonju University Wansan-gu Jeonju-si Jeollabuk-do Korea Agency for Defense Development Korea School of Mechanical Engineering Pusan National University Geumjeong-gu Busan Korea

ISBN: (纸本)9781509015740;9781509015733

The application of robust control method is required for the design of the automatic control system of the unmanned underwater vehicle which is operated in underwater environment under great disturbance. Sliding Perturbation Observer (SPO) was developed to overcome the chattering phenomenon which is a weak point of the sliding mode control. This control method can be applied to the control of underwater vehicle operating in strong environmental disturbances because it has the advantage of estimating and compensating for nonlinear components, modeling uncertainties and environmental disturbances of the system. In this study, the sliding mode control with sliding perturbation observer (SMCSPO) is simulated in an underwater vehicle to evaluate the disturbance compensation performance and compare it with a PID controller.

关键词： Underwater vehicles Perturbation methods Vehicle dynamics Observers Mathematical model Sliding mode control PI control

来源：评论

学校读者我要写书评

暂无评论

WaterGAN: Unsupervised generative network to enable real-time color correction of monocular underwater images

arXiv

引用

arXiv 2017年

作者： Li, Jie Skinner, Katherine A. Eustice, Ryan M. Johnson-Roberson, Matthew Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109 United States Robotics Program University of Michigan Ann ArborMI48109 United States Department of Naval Architecture and Marine Engineering University of Michigan Ann ArborMI48109 United States

This paper reports on WaterGAN, a generative adversarial network (GAN) for generating realistic underwater images from in-air image and depth pairings in an unsupervised pipeline used for color correction of monocular underwater images. Cameras onboard autonomous and remotely operated vehicles can capture high resolution images to map the seafloor;however, underwater image formation is subject to the complex process of light propagation through the water column. The raw images retrieved are characteristically different than images taken in air due to effects such as absorption and scattering, which cause attenuation of light at different rates for different wavelengths. While this physical process is well described theoretically, the model depends on many parameters intrinsic to the water column as well as the structure of the scene. These factors make recovery of these parameters difficult without simplifying assumptions or field calibration;hence, restoration of underwater images is a non-trivial problem. Deep learning has demonstrated great success in modeling complex nonlinear systems but requires a large amount of training data, which is difficult to compile in deep sea environments. Using WaterGAN, we generate a large training dataset of corresponding depth, in-air color images, and realistic underwater images. This data serves as input to a two-stage network for color correction of monocular underwater images. Our proposed pipeline is validated with testing on real data collected from both a pure water test tank and from underwater surveys collected in the field. Source code, sample datasets, and pretrained models are made publicly available. Copyright © 2017, The Authors. All rights reserved.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Driving in the Matrix: Can virtual worlds replace human-generated annotations for real world tasks?

Driving in the Matrix: Can virtual worlds replace human-gene...

引用

IEEE International Conference on robotics and Automation (ICRA)

作者： Matthew Johnson-Roberson Charles Barto Rounak Mehta Sharath Nittur Sridhar Karl Rosaen Ram Vasudevan Department of Naval Architecture and Marine Engineering University of Michigan Ann Arbor MI USA Electrical Engineering and Computer Science University of Michigan Ann Arbor MI USA Robotics Program University of Michigan Ann Arbor MI USA Mechanical Engineering University of Michigan Ann Arbor MI USA

关键词： Engines Training Machine learning Data models Robots Training data Automobiles

来源：评论

学校读者我要写书评

暂无评论

Gene expression imputation across multiple brain regions provides insights into schizophrenia risk (vol 51, pg 659, 2019)

引用

NATURE GENETICS 2019年第6期51卷 1068-1068页

作者： Huckins, Laura M. Dobbyn, Amanda Ruderfer, Douglas M. Hoffman, Gabriel Wang, Weiqing Pardinas, Antonio F. Rajagopal, Veera M. Als, Thomas D. Nguyen, Hoang T. Girdhar, Kiran Boocock, James Roussos, Panos Fromer, Menachem Kramer, Robin Domenici, Enrico Gamazon, Eric R. Purcell, Shaun Demontis, Ditte Borglum, Anders D. Walters, James T. R. O'Donovan, Michael C. Sullivan, Patrick Owen, Michael J. Devlin, Bernie Sieberts, Solveig K. Cox, Nancy J. Im, Hae Kyung Sklar, Pamela Stahl, Eli A. Pamela Sklar Division of Psychiatric Genomics Icahn School of Medicine at Mount Sinai New York NY USA Department of Genetics and Genomics Icahn School of Medicine at Mount Sinai New York NY USA Department of Psychiatry Icahn School of Medicine at Mount Sinai New York NY USA Icahn Institute for Genomics and Multiscale Biology Icahn School of Medicine at Mount Sinai New York NY USA Vanderbilt University Medical Center Nashville TN USA MRC Centre for Neuropsychiatric Genetics and Genomics Cardiff University Cardiff UK Department of Biomedicine Aarhus University Aarhus Denmark The Lundbeck Foundation Initiative for Integrative Psychiatric Research iPSYCH Denmark Center for Integrative Sequencing Aarhus University Aarhus Denmark Department of Human Genetics David Geffen School of Medicine University of California Los Angeles Los Angeles CA USA Human Brain Collection Core National Institute of Mental Health Bethesda MD USA Laboratory of Neurogenomic Biomarkers Centre for Integrative Biology (CIBIO) University of Trento Trento Italy Clare Hall University of Cambridge Cambridge UK University of North Carolina at Chapel Hill Chapel Hill NC USA Karolinska Institutet Stockholm Sweden Department of Psychiatry University of Pittsburgh Pittsburgh PA USA Systems Biology Sage Bionetworks Seattle WA USA Section of Genetic Medicine Department of Medicine University of Chicago Chicago IL USA Integrated Technology Research Laboratories Pharmaceutical Research Division Takeda Pharmaceutical Company Limited Fujisawa Japan Neuropsychiatry Section Department of Psychiatry Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Neuropsychiatric Signaling Program Department of Psychiatry Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Psychiatry JJ Peters Virginia Medical Center Bronx NY USA Department of Neuroscience Icahn School of Medicine at Mount Sinai New York New York NY USA Friedman Brain Institute

Transcriptomic imputation approaches combine eQTL reference panels with large-scale genotype data in order to test associations between disease and gene expression. These genic associations could elucidate signals in complex genome-wide association study (GWAS) loci and may disentangle the role of different tissues in disease development. We used the largest eQTL reference panel for the dorso-lateral prefrontal cortex (DLPFC) to create a set of gene expression predictors and demonstrate their utility. We applied DLPFC and 12 GTEx-brain predictors to 40,299 schizophrenia cases and 65,264 matched controls for a large transcriptomic imputation study of schizophrenia. We identified 413 genic associations across 13 brain regions. Stepwise conditioning identified 67 non-MHC genes, of which 14 did not fall within previous GWAS loci. We identified 36 significantly enriched pathways, including hexosaminidase-A deficiency, and multiple porphyric disorder pathways. We investigated developmental expression patterns among the 67 non-MHC genes and identified specific groups of pre- and postnatal expression.

关键词： Gene expression Genome-wide association studies Schizophrenia Transcriptomics

来源：评论

学校读者我要写书评

暂无评论

BioTIME 2.0: Expanding and Improving a Database of Biodiversity Time Series

引用

Global Ecology and Biogeography 2025年第5期34卷

作者： Maria Dornelas Laura H. Antão Amanda E. Bates Viviana Brambilla Jonathan M. Chase Cher F. Y. Chow Ada Fontrodona-Eslava Anne E. Magurran Inês S. Martins Faye Moyes Alban Sagouis Samuel Adu-Acheampong Daniel Acquah-Lamptey Dušan Adam Penelope A. Ajani Aitor Albaina Pablo Almaraz Jeongseop An Roger Sigismund Anderson Madelaine Jean Robertson Anderson Alexsander Z. Antunes Ivan Arismendi Linda Armbrecht Pedro Aros-Mardones Sreejith Kalpuzha Ashtamoorthy Narayanan Ayyappan Gal Badihi Joseph J. Bailey Andrew H. Baird Mark Edward Baird Sreekumar Vadakkethil Balakrishnan José António L. Barão-Nóbrega Adi Barash Miguel Barbosa Jos Barlow Claus Bässler Matthieu Beaumont Natalie Beenaerts Tiago Octavio Begot Wallace Beiroz Ricardo Beldade David M. Bell Alecia Bellgrove Jonathan Belmaker Lisandro Benedetti-Cecchi Cassandra E. Benkwitt Pamela Medina-van Berkum Brandon T. Bestelmeyer Matthew G. Betts Maxwell Kelvin Billah Anne D. Bjorkman Magdalena Błażewicz Christopher P. Bloch Shane A. Blowes Antonio Bode Juliano A. Bogoni Thomas Bolger Timothy C. Bonebrake Erik Bonsdorff Roberta Bottarin Luke N. Brokensha Rob W. Brooker Andrew J. Brooks Helge Bruelheide Thiago Almeida Bueno Claire Laguionie Mariana Lopes Campagnoli James Cant Erica Pellegrini Caramaschi Alexandre Caron Tadhg Carroll Tancredi Caruso Juan Carvajal-Quintero Giuseppe Castaldelli Edward Castañeda-Moya Pedro V. Castilho Sonia Zanini Cechin Shahar Chaikin Uchangi Manjunatha Chandrashekara Tory J. Chase Chaolun Allen Chen Jorge José Cherem Sei-Woong Choi Erica M. Christensen Alexander V. Christianini Jackson Wing Four Chu Peter Coad Carl Van Colen Lise Comte Elisabeth J. Cooper J. Hans C. Cornelissen Eddy Cosson Unai Cotano Luc Crevecoeur Shannan Kyle Crow Graeme S. Cumming Vanessa S. Daga Gabriella Damasceno Gergana N. Daskalova Claire H. Davies Robert A. Davis Frank P. Day Sussy De-La-Zerda Amy Elizabeth Deacon Indradatta de Castro-Arrazola Steven Degraer Kharran Deonarinesingh Juan C. Diaz-Ricaurte Christopher R. Dickman Tara Dirilgen Ciaran John Dolan J. Emmet Centre for Biological Diversity School of Biology University of St Andrews Fife UK MARE—Centro de Ciências do Mar e do Ambiente Faculdade de Ciências Universidade de Lisboa Lisboa Portugal Research Centre for Ecological Change Organismal and Evolutionary Biology Research Programme Faculty of Biological and Environmental Sciences University of Helsinki Helsinki Finland Department of Biology University of Victoria Victoria British Columbia Canada German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig Leipzig Germany Department of Computer Science Martin Luther University Halle-Wittenberg Halle (Saale) Germany Leverhulme Centre for Anthropocene Biodiversity Department of Biology University of York York UK Department of Animal Biology and Conservation Science University of Ghana Accra Ghana Philipps Universität Marburg Marburg Germany Department of Forest Ecology Silva Tarouca Research Institute Brno Czech Republic School of Life Sciences University of Technology Sydney Sydney New South Wales Australia Zoology and Animal Cell Biology Department University of the Basque Country (UPV/EHU) Leioa Spain Grupo de Oceanografía de Ecosistemas Instituto de Ciencias Marinas de Andalucía CSIC Campus Universitario de Puerto Real Puerto Real Spain National Institute of Ecology Seocheon Republic of Korea African Regional Postgraduate Programme in Insect Science University of Ghana Accra Ghana Département de Biologie Université de Sherbrooke Sherbrooke Quebec Canada Núcleo de Conservação da Biodiversidade Instituto de Pesquisas Ambientais Secretaria de Meio Ambiente Infraestrutura e Logística do Estado de São Paulo São Paulo Brazil Department of Fisheries Wildlife and Conservation Sciences Oregon State University Corvallis Oregon USA Institute for Marine and Antarctic Studies Ecology & Biodiversity Centre University of Tasmania Hobart Tasmania Australia Graduate Program in Oceanography Department of Oceanography Faculty of Natural Scienc

Motivation Here, we make available a second version of the BioTIME database, which compiles records of abundance estimates for species in sample events of ecological assemblages through time. The updated version expands version 1.0 of the database by doubling the number of studies and includes substantial additional curation to the taxonomic accuracy of the records, as well as the metadata. Moreover, we now provide an R package (BioTIMEr) to facilitate use of the database. Main Types of Variables Included The database is composed of one main data table containing the abundance records and 11 metadata tables. The data are organised in a hierarchy of scales where 11,989,233 records are nested in 1,603,067 sample events, from 553,253 sampling locations, which are nested in 708 studies. A study is defined as a sampling methodology applied to an assemblage for a minimum of 2 years. Spatial Location and Grain Sampling locations in BioTIME are distributed across the planet, including marine, terrestrial and freshwater realms. Spatial grain size and extent vary across studies depending on sampling methodology. We recommend gridding of sampling locations into areas of consistent size. Time Period and Grain The earliest time series in BioTIME start in 1874, and the most recent records are from 2023. Temporal grain and duration vary across studies. We recommend doing sample-level rarefaction to ensure consistent sampling effort through time before calculating any diversity metric. Major Taxa and Level of Measurement The database includes any eukaryotic taxa, with a combined total of 56,400 taxa. Software Format csv and. SQL.

关键词： assemblage biodiversity community database ecoinformatics species taxa time

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：