检索结果-内蒙古大学图书馆

The IDEAL framework for surgical robotics: development, comparative evaluation and long-term monitoring (vol 30, pg 61, 2024)

引用

NATURE MEDICINE 2024年第4期30卷 1213-1213页

作者： Marcus, Hani J. Ramirez, Pedro T. Khan, Danyal Z. Horsfall, Hugo Layard Hanrahan, John G. Williams, Simon C. Beard, David J. Bhat, Rani Catchpole, Ken Cook, Andrew Hutchison, Katrina Martin, Janet Melvin, Tom Stoyanov, Danail Rovers, Maroeska Raison, Nicholas Dasgupta, Prokar Noonan, David Stocken, Deborah Sturt, Georgia Vanhoestenberghe, Anne Vasey, Baptiste McCulloch, Peter Department of Neurosurgery National Hospital of Neurology and Neurosurgery London UK Wellcome/Engineering and Physical Sciences Research Council (EPSRC) Centre for Interventional and Surgical Sciences (WEISS) London UK Department of Obstetrics and Gynaecology Houston Methodist Hospital Neal Cancer Center Houston TX USA RCS Surgical Interventional Trials Unit (SITU) & Robotic and Digital Surgery Initiative (RADAR) Nuffield Dept Orthopaedics Rheumatology and Musculo-skeletal Sciences University of Oxford Oxford UK Department of Gynaecological Oncology Apollo Hospital Bengaluru India Department of Anaesthesia and Perioperative Medicine Medical University of South Carolina Charleston SC USA NIHR Coordinating Centre and Clinical Trials Unit University of Southampton Southampton UK Department of Philosophy Macquarie University Sydney Australia Department of Anesthesia & Perioperative Medicine University of Western Ontario Ontario Canada Department of Medical Gerontology School of Medicine Trinity College Dublin Dublin Republic of Ireland Department of Medical Imaging Radboudumc Nijmegen the Netherlands Department of Biomedical Engineering and Imaging Sciences King’s College London London UK King’s Health Partners Academic Surgery King’s College London London UK Moon Surgical Paris France RCSEng Surgical Trials Centre Leeds Institute of Clinical Trials Research University of Leeds Leeds UK Bowel Research UK London UK School of Biomedical Engineering & Imaging Sciences King’s College London London UK Department of Surgery Geneva University Hospital Geneva Switzerland Nuffield Department of Surgical Sciences University of Oxford John Radcliffe Hospital Oxford UK PRISMA Lab University of Naples Federico II Naples Italy Health Ethics and Policy Lab Department of Health Sciences and Technology Zürich Switzerland School of Computer Science University of Birmingham Birmingham UK Division of Cardiac Surgery Brigham and Women’s Hospital Harvard Medical Schoo

The next generation of surgical robotics is poised to disrupt healthcare systems worldwide, requiring new frameworks for evaluation. However, evaluation during a surgical robot’s development is challenging due to their complex evolving nature, potential for wider system disruption and integration with complementary technologies like artificial intelligence. Comparative clinical studies require attention to intervention context, learning curves and standardized outcomes. Long-term monitoring needs to transition toward collaborative, transparent and inclusive consortiums for real-world data collection. Here, the Idea, Development, Exploration, Assessment and Long-term monitoring (IDEAL) robotics Colloquium proposes recommendations for evaluation during development, comparative study and clinical monitoring of surgical robots—providing practical recommendations for developers, clinicians, patients and healthcare systems. Multiple perspectives are considered, including economics, surgical training, human factors, ethics, patient perspectives and sustainability. Further work is needed on standardized metrics, health economic assessment models and global applicability of recommendations.

关键词： Medical research Surgery

来源：评论

学校读者我要写书评

暂无评论

MutualNet: Adaptive ConvNet via Mutual Learning from Different Model Configurations

arXiv

引用

arXiv 2021年

作者： Yang, Taojiannan Zhu, Sijie Mendieta, Matias Wang, Pu Balakrishnan, Ravikumar Lee, Minwoo Han, Tao Shah, Mubarak Chen, Chen The Center for Research in Computer Vision University of Central Florida OrlandoFL32816 United States The Department of Computer Science University of North Carolina at Charlotte CharlotteNC28223 United States Intel Labs New Jersey Institute of Technology NewarkNJ07102 United States The Department of Electrical and Computer Engineering New Jersey Institute of Technology NewarkNJ07102 United States

Most existing deep neural networks are static, which means they can only perform inference at a fixed complexity. But the resource budget can vary substantially across different devices. Even on a single device, the affordable budget can change with different scenarios, and repeatedly training networks for each required budget would be incredibly expensive. Therefore, in this work, we propose a general method called MutualNet to train a single network that can run at a diverse set of resource constraints. Our method trains a cohort of model configurations with various network widths and input resolutions. This mutual learning scheme not only allows the model to run at different width-resolution configurations but also transfers the unique knowledge among these configurations, helping the model to learn stronger representations overall. MutualNet is a general training methodology that can be applied to various network structures (e.g., 2D networks: MobileNets, ResNet, 3D networks: SlowFast, X3D) and various tasks (e.g., image classification, object detection, segmentation, and action recognition), and is demonstrated to achieve consistent improvements on a variety of datasets. Since we only train the model once, it also greatly reduces the training cost compared to independently training several models. Surprisingly, MutualNet can also be used to significantly boost the performance of a single network, if dynamic resource constraints are not a concern. In summary, MutualNet is a unified method for both static and adaptive, 2D and 3D networks. Code and pre-trained models are available at https://***/taoyang1122/MutualNet. Copyright © 2021, The Authors. All rights reserved.

关键词： Budget control

来源：评论

学校读者我要写书评

暂无评论

Bacteria Flagella-Mimicking Polymer Multilayer Magnetic Microrobots

SSRN

引用

SSRN 2024年

作者： Wu, Zhiguang Lu, Liang Bai, Shuang Shi, Jiaqi Zhang, Zhuotao Pan, Yunlu Wang, Wei Wu, Wenfeng Sun, Shoubin Li, Xu Hou, Gang Huang, Tianyun Jia, Yuxin Zuo, Kaiwen Granovsky, Alexander Wu, He Nikolai, Perov School of Medicine and Health Harbin Institute of Technology Harbin China State Key Laboratory of Robotics and System Harbin Institute of Technology Harbin China The First Affiliated Hospital of Harbin Medical University Harbin China National Center for Respiratory Medicine State Key Laboratory of Respiratory Health and Multimorbidity National Clinical Research Center for Respiratory Diseases Institute of Respiratory Medicine Chinese Academy of Medical Sciences Department of Pulmonary and Critical Care Medicine Center of Respiratory Medicine China-Japan Friendship Hospital Beijing China College of Science Sichuan Agricultural University Chengdu China The Fourth Affiliated Hospital of Harbin Medical University Harbin China Department of Ophthalmology The Second Hospital Jilin University Changchun China Department of Advanced Manufacturing and Robotics College of Engineering State Key Laboratory for Turbulence and Complex Systems Peking University Beijing China Department of Mathematics Imperial College London United Kingdom Department of Computer Science University of Warwick Coventry United Kingdom Magnetism Department Faculty of Physics Lomonosov Moscow State University Moscow Russia Harbin Medical University China

Learning from natural bacteria flagellum, we demonstrate a magnetic polymer multilayer conical microrobot that bestow the controllable propulsion upon external rotating magnetic field with uniform intensity. The magnetic microrobots were constructed by template-assisted Layer-by-Layer technique technique and subsequent functionalization of magnetic particle onto the big opening of the microrobots. Geometric variables of the polymer microrobots, such as the diameter and wall thickness, can be controlled by selection of porous template and layers of assembly. The microrobots perform controllable propulsion through the manipulation of magnetic field. The comparative analysis of the movement behavior reveals that the deformation of microrobots may attributed to the propulsion upon rotating magnetic field, which is similar as that of natural bacteria. The influence of actuation frequencies and frequencies on the velocity of the microrobots is studied. Such use of polymer polymer multilayer tubular micro/nanoarchitectures results in significant savings in the processing costs and provides a simple and cost-effective fabrication route for the mass production of magnetic polymer microrobots. © 2024, The Authors. All rights reserved.

关键词： Bacteria

来源：评论

学校读者我要写书评

暂无评论

OTE: Optimal Trustworthy EdgeAI solutions for smart cities

OTE: Optimal Trustworthy EdgeAI solutions for smart cities

引用

IEEE/ACM International Symposium on Cluster Computing and the Grid (CCGRID)

作者： Vasileios Mygdalis Lorenzo Carnevale Jose Ramiro Martí nez-De-Dios Dmitriy Shutin Giovanni Aiello Massimo Villari Ioannis Pitas Department of Informatics Aristotle University of Thessaloniki Thessaloniki Greece Department of Mathematical and Computer Science Physics and Hearth Sciences University of Messina Messina Italy Gruppo Nazionale per il Calcolo Scientifico (GNCS) Istituto Nazionale di Alta Matematica (INdAM) &#x201C F. Severi&#x201D Rome Italy Robotics Vision and Control Group University of Seville Seville Spain Institute of Communications and Navigation German Aerospace Center (DLR) Wessling Germany Research and Development Lab. Engineering Ingegneria Informatica S.p.A Rome Italy

This work studies and defines the problem of providing extensive and opportunistic Edge AI-based area coverage in smart city application scenarios, by researching and determining the optimal configuration of sensing and computational resources for minimizing the environmental/technology footprint of the solution. A typical smart city computing continuum consists of statically installed multimodal sensing Internet-of-Things (IoT) nodes at various city locations, accompanied by interconnected computational Cloud/Edge/IoT nodes. This paper presents Optimal Trustworthy EdgeAI (OTE), an entirely novel research pipeline, that complements existing smart city infrastructure with intelligent drone Edge/IoT nodes (in the form of modularly equipped unmanned aerial vehicles), capable of autonomous repositioning according to individual/collective sensing and coverage criteria. Thereby, we envisage the emerging cutting-edge technologies of trustworthy sensing, perceiving, modelling technologies for predicting the behavior of moving targets (e.g., citizens/vehicles/objects), understanding natural phenomena (e.g., sea wave motion, urban flora/fauna, biodiversity) in order to anticipate events (people's bad habits, environmental changes), by exploiting novel continuous data processing services across the whole span of the enhanced Cloud-Edge-IoT computing continuum.

关键词： Cloud computing Privacy Smart cities Pipelines Semantics Robot sensing systems Software

来源：评论

学校读者我要写书评

暂无评论

Opportunities and challenges for monitoring terrestrial biodiversity in the robotics age

引用

Nature Ecology and Evolution 2025年第6期9卷 1031-1042页

作者： Pringle, Stephen Dallimer, Martin Goddard, Mark A. Le Goff, Léni K. Hart, Emma Langdale, Simon J. Fisher, Jessica C. Abad, Sara-Adela Ancrenaz, Marc Angeoletto, Fabio Auat Cheein, Fernando Austen, Gail E. Bailey, Joseph J. Baldock, Katherine C. R. Banin, Lindsay F. Banks-Leite, Cristina Barau, Aliyu S. Bashyal, Reshu Bates, Adam J. Bicknell, Jake E. Bielby, Jon Bosilj, Petra Bush, Emma R. Butler, Simon J. Carpenter, Dan Clements, Christopher F. Cully, Antoine Davies, Kendi F. Deere, Nicolas J. Dodd, Michael Drinkwater, Rosie Driscoll, Don A. Dutilleux, Guillaume Dyrmann, Mads Edwards, David P. Farhadinia, Mohammad S. Faruk, Aisyah Field, Richard Fletcher, Robert J. Foster, Chris W. Fox, Richard Francksen, Richard M. Franco, Aldina M. A. Gainsbury, Alison M. Gardner, Charlie J. Giorgi, Ioanna Griffiths, Richard A. Hamaza, Salua Hanheide, Marc Hayward, Matt W. Hedblom, Marcus Helgason, Thorunn Heon, Sui P. Hughes, Kevin A. Hunt, Edmund R. Ingram, Daniel J. Jackson-Mills, George Jowett, Kelly Keitt, Timothy H. Kloepper, Laura N. Kramer-Schadt, Stephanie Labisko, Jim Labrosse, Frédéric Lawson, Jenna Lecomte, Nicolas de Lima, Ricardo F. Littlewood, Nick A. Marshall, Harry H. Masala, Giovanni L. Maskell, Lindsay C. Matechou, Eleni Mazzolai, Barbara McConnell, Alistair Melbourne, Brett A. Miriyev, Aslan Nana, Eric Djomo Ossola, Alessandro Papworth, Sarah Parr, Catherine L. Payo-Payo, Ana Perry, Gad Pettorelli, Nathalie Pillay, Rajeev Potts, Simon G. Prendergast-Miller, Miranda T. Qie, Lan Rolley-Parnell, Persie Rossiter, Stephen J. Rowcliffe, Marcus Rumble, Heather Sadler, Jon P. Sandom, Christopher J. Sanyal, Asiem Schrodt, Franziska Sethi, Sarab S. Shabrani, Adi Siddall, Robert Smith, Simón C. Snep, Robbert P. H. Soulsbury, Carl D. Stanley, Margaret C. Stephens, Philip A. Stephenson, P.J. Struebig, Matthew J. Studley, Matthew Svátek, Martin Tang, Gilbert Taylor, Nicholas K. Umbers, Kate D. L. Ward, Robert J. White, Patrick J. C. Whittingham, Mark J. Wich, Serge Williams, Christopher D. Yakubu, Ibrahim B. Yoh, Natalie Durrell Institute of Conservation and Ecology (DICE) School of Natural Sciences University of Kent Canterbury United Kingdom Centre for Environmental Policy Imperial College London London United Kingdom Department of Geography and Environmental Sciences Northumbria University Newcastle upon Tyne United Kingdom School of Computing Engineering and the Built Environment Edinburgh Napier University Edinburgh United Kingdom Synthotech Ltd Milner Court Hornbeam Square Harrogate United Kingdom Mechanical Engineering Department University College London London United Kingdom HUTAN SWD Kota Kinabalu Malaysia Programa de Pós-Graduação em Gestão e Tecnologia Ambiental da Universidade Federal de Rondonópolis Rondonópolis Brazil Department of Engineering Harper Adams University Newport United Kingdom Applied Ecology Research Group School of Life Sciences Anglia Ruskin University Cambridge United Kingdom Centre for Ecology and Hydrology Penicuik United Kingdom Department of Life Sciences Imperial College London Silwood Park Campus Ascot United Kingdom Department of Urban and Regional Planning Faculty of Earth and Environmental Sciences Bayero University Kano Nigeria Greenhood Nepal Kathmandu Nepal Animal Rural & Environmental Sciences Nottingham Trent University Nottinghamshire United Kingdom School of Biological and Environmental Sciences Liverpool John Moores University Liverpool United Kingdom Lincoln Centre for Autonomous Systems University of Lincoln Lincoln United Kingdom Royal Botanic Garden Edinburgh Edinburgh United Kingdom School of Biological Sciences University of East Anglia Norwich Research Park Norwich United Kingdom Digital Ecology Limited Bristol United Kingdom School of Biological Sciences University of Bristol Bristol United Kingdom Department of Computing Imperial College London London United Kingdom Department of Ecology and Evolutionary Biology University of Colorado Boulder CO United States Faculty of Science Technology Engineering

With biodiversity loss escalating globally, a step change is needed in our capacity to accurately monitor species populations across ecosystems. Robotic and autonomous systems (RAS) offer technological solutions that may substantially advance terrestrial biodiversity monitoring, but this potential is yet to be considered systematically. We used a modified Delphi technique to synthesize knowledge from 98 biodiversity experts and 31 RAS experts, who identified the major methodological barriers that currently hinder monitoring, and explored the opportunities and challenges that RAS offer in overcoming these barriers. Biodiversity experts identified four barrier categories: site access, species and individual identification, data handling and storage, and power and network availability. robotics experts highlighted technologies that could overcome these barriers and identified the developments needed to facilitate RAS-based autonomous biodiversity monitoring. Some existing RAS could be optimized relatively easily to survey species but would require development to be suitable for monitoring of more ‘difficult’ taxa and robust enough to work under uncontrolled conditions within ecosystems. Other nascent technologies (for instance, new sensors and biodegradable robots) need accelerated research. Overall, it was felt that RAS could lead to major progress in monitoring of terrestrial biodiversity by supplementing rather than supplanting existing methods. Transdisciplinarity needs to be fostered between biodiversity and RAS experts so that future ideas and technologies can be codeveloped effectively. © The Author(s) 2025.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Capsule vision 2024 Challenge: Multi-Class Abnormality Classification for Video Capsule Endoscopy

arXiv

引用

arXiv 2024年

作者： Handa, Palak Mahbod, Amirreza Schwarzhans, Florian Woitek, Ramona Goel, Nidhi Dhir, Manas Chhabra, Deepti Jha, Shreshtha Sharma, Pallavi Thakur, Vijay Chawla, Simarpreet Singh Gunjan, Deepak Kakarla, Jagadeesh Raman, Balasubramanian Research Center for Medical Image Analysis and Artificial Intelligence Department of Medicine Danube Private University Krems Austria Department of Electronics and Communication Engineering Indira Gandhi Delhi Technical University for Women Delhi India Department of Artificial Intelligence and Data Sciences Indira Gandhi Delhi Technical University for Women Delhi India Department of Artificial Intelligence and Machine Learning University School of Automation and Robotics Guru Gobind Singh Indraprastha University Delhi India Department of Electronics and Communication Engineering Delhi Technological University Delhi India Columbia University New YorkNY United States Department of Gastroenterology and HNU All India Institute of Medical Sciences Delhi India Chennai Kancheepuram India Department of Computer Science and Engineering Indian Institute of Technology Roorkee India

We present the Capsule vision 2024 Challenge: Multi-Class Abnormality Classification for Video Capsule Endoscopy. It was virtually organized by the research center for Medical Image Analysis and Artificial Intelligence (MIAAI), department of Medicine, Danube Private University, Krems, Austria in collaboration with the 9th International Conference on computer vision & Image Processing (CVIP 2024) being organized by the Indian Institute of Information Technology, Design and Manufacturing (IIITDM) Kancheepuram, Chennai, India. This document provides an overview of the challenge, including the registration process, rules, submission format, description of the datasets used, qualified team rankings, all team descriptions, and the benchmarking results reported by the organizers. © 2024, CC BY.

关键词： Endoscopy

来源：评论

学校读者我要写书评

暂无评论

Multi-Class Segmentation of Aortic Branches and Zones in Computed Tomography Angiography: The AortaSeg24 Challenge

arXiv

引用

arXiv 2025年

作者： Imran, Muhammad Krebs, Jonathan R. Sivaraman, Vishal Balaji Zhang, Teng Kumar, Amarjeet Ueland, Walker R. Fassler, Michael J. Huang, Jinlong Sun, Xiao Wang, Lisheng Shi, Pengcheng Rokuss, Maximilian Baumgartner, Michael Kirchhof, Yannick Maier-Hein, Klaus H. Isensee, Fabian Liu, Shuolin Han, Bing Nguyen, Bong Thanh Shin, Dong-Jin Ji-Woo, Park Choi, Mathew Uhm, Kwang-Hyun Ko, Sung-Jea Lee, Chanwoong Chun, Jaehee Kim, Jin Sung Zhang, Minghui Zhang, Hanxiao You, Xin Gu, Yun Pan, Zhaohong Liu, Xuan Liang, Xiaokun Tiefenthaler, Markus Almar-Munoz, Enrique Schwab, Matthias Kotyushev, Mikhail Epifanov, Rostislav Wodzinski, Marek Müller, Henning Qayyum, Abdul Mazher, Moona Niederer, Steven A. Wang, Zhiwei Yang, Kaixiang Ren, Jintao Korreman, Stine Sofia Gao, Yuchong Zeng, Hongye Zheng, Haoyu Zheng, Rui Yue, Jinghua Zhou, Fugen Liu, Bo Cosman, Alexander Liang, Muxuan Zhao, Chang Upchurch, Gilbert R. Ma, Jun Zhou, Yuyin Cooper, Michol A. Shao, Wei Department of Medicine University of Florida GainesvilleFL32611 United States Department of Surgery University of Florida GainesvilleFL32611 United States Department of Electrical and Computer Engineering University of Florida GainesvilleFL32611 United States Department of Computer Information Science and Engineering University of Florida GainesvilleFL32611 United States Shanghai Jiao Tong University Shanghai China Electronic & Information Engineering School Harbin Institute of Technology [Shenzhen Shenzhen China Division of Medical Image Computing German Cancer Research Center [DKFZ Heidelberg Germany CO. LTD Beijing China MedAI Seoul Korea Republic of Korea University Seoul Korea Republic of Department of Radiation Oncology Yonsei Cancer Center Heavy Ion Therapy Research Institute Yonsei University College of Medicine Seoul Korea Republic of Institute of Medical Robotics Shanghai Jiao Tong University Shanghai China Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China Medical University of Innsbruck Innsbruck Austria Novosibirsk State University Novosibirsk Russia AGH University of Krakow Krakow Poland Sierre Switzerland University of Geneva Geneva Switzerland The Sense Innovation and Research Center Sion Switzerland National Heart and Lung Institute Faculty of Medicine Imperial College London London United Kingdom Wuhan National Laboratory for Optoelectronics Huazhong University of Science and Technology Hubei Wuhan China Aarhus University Department of Clinical Medicine Nordre Palle Juul-Jensens Blvd. 11 Aarhus8200 Denmark Shanghaitech University Pudong Shanghai China Image Processing Center Beihang University Beijing China Department of Computer Science and Engineering University of California Santa CruzCA95064 United States Department of Biostatistics University of Florida GainesvilleFL32611 United States University Health Network ON Canada Agronomy Department University of F

Multi-class segmentation of the aorta in computed tomography angiography (CTA) scans is essential for diagnosing and planning complex endovascular treatments for patients with aortic dissections. However, existing methods reduce aortic segmentation to a binary problem, limiting their ability to measure diameters across different branches and zones. Furthermore, no open-source dataset is currently available to support the development of multi-class aortic segmentation methods. To address this gap, we organized the AortaSeg24 MICCAI Challenge, introducing the first dataset of 100 CTA volumes annotated for 23 clinically relevant aortic branches and zones. This dataset was designed to facilitate both model development and validation. The challenge attracted 121 teams worldwide, with participants leveraging state-of-the-art frameworks such as nnU-Net and exploring novel techniques, including cascaded models, data augmentation strategies, and custom loss functions. We evaluated the submitted algorithms using the Dice Similarity Coefficient (DSC) and Normalized Surface Distance (NSD), highlighting the approaches adopted by the top five performing teams. This paper presents the challenge design, dataset details, evaluation metrics, and an in-depth analysis of the top-performing algorithms. The annotated dataset, evaluation code, and implementations of the leading methods are publicly available to support further research. All resources can be accessed at https://***. © 2025, CC BY-NC-SA.

关键词： computerized tomography

来源：评论

学校读者我要写书评

暂无评论

Generative Pseudo-label Refinement for Unsupervised Domain Adaptation

arXiv

引用

arXiv 2020年

作者： Morerio, Pietro Volpi, Riccardo Ragonesi, Ruggero Murino, Vittorio Pattern Analysis & Computer Vision - Istituto Italiano di Tecnologia Computer Science Department Università di Verona Italy Huawei Technologies Ltd. Ireland Research Center

We investigate and characterize the inherent resilience of conditional Generative Adversarial Networks (cGANs) against noise in their conditioning labels, and exploit this fact in the context of Unsupervised Domain Adaptation (UDA). In UDA, a classifier trained on the labelled source set can be used to infer pseudo-labels on the unlabelled target set. However, this will result in a significant amount of misclassified examples (due to the well-known domain shift issue), which can be interpreted as noise injection in the ground-truth labels for the target set. We show that cGANs are, to some extent, robust against such "shift noise". Indeed, cGANs trained with noisy pseudo-labels, are able to filter such noise and generate cleaner target samples. We exploit this finding in an iterative procedure where a generative model and a classifier are jointly trained: in turn, the generator allows to sample cleaner data from the target distribution, and the classifier allows to associate better labels to target samples, progressively refining target pseudo-labels. Results on common benchmarks show that our method performs better or comparably with the unsupervised domain adaptation state of the art. Copyright © 2020, The Authors. All rights reserved.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Cassandra: Detecting trojaned networks from adversarial perturbations

arXiv

引用

arXiv 2020年

作者： Zhang, Xiaoyu Mian, Ajmal Gupta, Rohit Rahnavard, Nazanin Shah, Mubarak Center for Research in Computer Vision University of Central Florida United States Computer Science and Software Engineering University of Western Australia Australia Department of Electrical and Computer Engineering University of Central Florida United States

Deep neural networks are being widely deployed for many critical tasks due to their high classification accuracy. In many cases, pre-trained models are sourced from vendors who may have disrupted the training pipeline to insert Trojan behaviors into the models. These malicious behaviors can be triggered at the adversary’s will and hence, cause a serious threat to the widespread deployment of deep models. We propose a method to verify if a pre-trained model is Trojaned or benign. Our method captures fingerprints of neural networks in the form of adversarial perturbations learned from the network gradients. Inserting backdoors into a network alters its decision boundaries which are effectively encoded in their adversarial perturbations. We train a two stream network for Trojan detection from its global (L∞ and L2 bounded) perturbations and the localized region of high energy within each perturbation. The former encodes decision boundaries of the network and latter encodes the unknown trigger shape. We also propose an anomaly detection method to identify the target class in a Trojaned network. Our methods are invariant to the trigger type, trigger size, training data and network architecture. We evaluate our methods on MNIST, NIST-Round0 and NIST-Round1 datasets, with up to 1,000 pre-trained models making this the largest study to date on Trojaned network detection, and achieve over 92% detection accuracy to set the new state-of-the-art. Copyright © 2020, The Authors. All rights reserved.

关键词： Encoding (symbols)

来源：评论

学校读者我要写书评

暂无评论

Unsupervised difficulty estimation with action scores

arXiv

引用

arXiv 2020年

作者： Arriaga, Octavio Valdenegro-Toro, Matias Department of Computer Science University of Bremen Bremen28359 Germany Robotics Innovation Center German Research Center for Artificial Intelligence Bremen28359 Germany

Evaluating difficulty and biases in machine learning models has become of extreme importance as current models are now being applied in real-world situations. In this paper we present a simple method for calculating a difficulty score based on the accumulation of losses for each sample during training. We call this the action score. Our proposed method does not require any modification of the model neither any external supervision, as it can be implemented as callback that gathers information from the training process. We test and analyze our approach in two different settings: image classification, and object detection, and we show that in both settings the action score can provide insights about model and dataset biases. Copyright © 2020, The Authors. All rights reserved.

关键词： Object detection

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：