检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Yuan, Kai Sajid, Noor Friston, Karl Li, Zhibin Embodied AI Lab Intel Labs Germany Wellcome Centre for Human Neuroimaging Queen Square Institute of Neurology University College London United Kingdom Department of Computer Science UCL Robotics Institute University College London United Kingdom

Humans can produce complex whole-body motions when interacting with their surroundings, by planning, executing and combining individual limb movements. We investigated this fundamental aspect of motor control in the setting of autonomous robotic operations. We approach this problem by hierarchical generative modelling equipped with multi-level planning—for autonomous task completion—that mimics the deep temporal architecture of human motor control. Here, temporal depth refers to the nested time scales at which successive levels of a forward or generative model unfold, for example, delivering an object requires a global plan to contextualise the fast coordination of multiple local movements of limbs. This separation of temporal scales also motivates robotics and control. Specifically, to achieve versatile sensorimotor control, it is advantageous to hierarchically structure the planning and low-level motor control of individual limbs. We use numerical and physical simulation to conduct experiments and to establish the efficacy of this formulation. Using a hierarchical generative model, we show how a humanoid robot can autonomously complete a complex task that necessitates a holistic use of locomotion, manipulation, and grasping. Specifically, we demonstrate the ability of a humanoid robot that can retrieve and transport a box, open and walk through a door to reach the destination, approach and kick a football, while showing robust performance in presence of body damage and ground irregularities. Our findings demonstrated the effectiveness of using human-inspired motor control algorithms, and our method provides a viable hierarchical architecture for the autonomous completion of challenging goal-directed tasks. © 2023, CC BY.

关键词： Anthropomorphic robots

来源：评论

学校读者我要写书评

暂无评论

Exploring Large Scale Pre-Trained Models for Robust Machine Anomalous Sound Detection

Exploring Large Scale Pre-Trained Models for Robust Machine ...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Bing Han Zhiqiang Lv Anbai Jiang Wen Huang Zhengyang Chen Yufeng Deng Jiawei Ding Cheng Lu Wei-Qiang Zhang Pingyi Fan Jia Liu Yanmin Qian Department of Computer Science and Engineering Auditory Cognition and Computational Acoustics Lab Shanghai Jiao Tong University Shanghai China Huakong AI Plus Company Limited Beijing China Department of Electronic Engineering Tsinghua University Beijing China School of Economics and Management North China Electric Power University Beijing China

Machine anomalous sound detection is a useful technique for various applications, but it often suffers from poor generalization due to the challenges of data collection and complex acoustic environment. To address this issue, we propose a robust machine anomalous sound detection model that leverages self-supervised pre-trained models on large-scale speech data. Specifically, we assign different weights to the features from different layers of the pre-trained model and then use the working condition as the label for self-supervised classification fine-tuning. Moreover, we introduce a data augmentation method that simulates different operating states of the machine to enrich the dataset. Furthermore, we devise a transformer pooling method that fuses the features of different segments. Experiments on the DCASE2023 dataset show that our proposed method outperforms the commonly used reconstruction-based autoencoder and classification-based convolutional network by a large margin, demonstrating the effectiveness of large-scale pre-training for enhancing the generalization and robustness of machine anomalous sound detection. In Task2 of DCASE2023, we achieve 2nd place with these methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Applying Pre-trained Multilingual BERT in Embeddings for Improved Malicious Prompt Injection Attacks Detection

arXiv

引用

arXiv 2024年

作者： Rahman, Md Abdur Shahriar, Hossain Wu, Fan Cuzzocrea, Alfredo Dept. of Intelligent Systems and Robotics University of West Florida PensacolaFL United States Center for Cybersecurity University of West Florida PensacolaFL United States Department of Computer Science Tuskegee University TuskegeeAL United States iDEA Lab University of Calabria Rende Italy

Large language models (LLMs) are renowned for their exceptional capabilities, and applying to a wide range of applications. However, this widespread use brings significant vulnerabilities. Also, it is well observed that there are huge gap which lies in the need for effective detection and mitigation strategies against malicious prompt injection attacks in large language models, as current approaches may not adequately address the complexity and evolving nature of these vulnerabilities in real-world applications. Therefore, this work focuses the impact of malicious prompt injection attacks which is one of most dangerous vulnerability on real LLMs applications. It examines to apply various BERT (Bidirectional Encoder Representations from Transformers) like multilingual BERT, DistilBert for classifying malicious prompts from legitimate prompts. Also, we observed how tokenizing the prompt texts and generating embeddings using multilingual BERT contributes to improve the performance of various machine learning methods: Gaussian Naive Bayes, Random Forest, Support Vector Machine, and Logistic Regression. The performance of each model is rigorously analyzed with various parameters to improve the binary classification to discover malicious prompts. Multilingual BERT approach to embed the prompts significantly improved and outperformed the existing works and achieves an outstanding accuracy of 96.55% by Logistic regression. Additionally, we investigated the incorrect predictions of the model to gain insights into its limitations. The findings can guide researchers in tuning various BERT for finding the most suitable model for diverse LLMs vulnerabilities. © 2024, CC BY.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Colibri5: Real-Time Monocular 5-DoF Trocar Pose Tracking for Robot-Assisted Vitreoretinal Surgery

Colibri5: Real-Time Monocular 5-DoF Trocar Pose Tracking for...

引用

IEEE International Conference on robotics and Automation (ICRA)

作者： Shervin Dehghani Michael Sommersperger Mahdi Saleh Alireza Alikhani Benjamin Busam Peter Gehlbach Iulian Iordachita Nassir Navab M. Ali Nasseri Department of Computer Science Technische Universität München München Germany Augenklinik und Poliklinik Klinikum Rechts der Isar der Technische Universität München München Germany Wilmer Eye Institute Johns Hopkins Hospital Baltimore MD USA Laboratory for Computational Sensing and Robotics Johns Hopkins University Baltimore MD USA

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

Retinal surgery is a complex medical procedure that requires high precision dexterity to perform delicate instrument maneuvers with sub-millimeter accuracy. Minimizing the manual tremor and achieving precise and repeatable execution of surgical tasks has motivated the development of robotic platforms to overcome the limitations of manual surgery. However, specific tasks, such as instrument insertion through the trocar, are more challenging in robotic surgery than in conventional manual procedures since the robot control is often optimized for navigation inside the eye. This challenges the integration of robotic systems, creating a high cognitive load on the operator and prolonging the surgery time. Moreover, misalignment of the robot’s remote center of motion (RCM) and trocar position during the procedure can lead to excessive forces between the instrument and the trocar, potentially causing patient trauma. Precise and rapid localization of the trocars enables the automation of the insertion procedure and dynamic compensation of eye *** this work, we present a real-time marker-less method for 3D pose tracking of trocar, achieved with only a single monocular camera. Our experiments show promising results towards real-time trocar pose estimation and tracking, achieving an average error of 3 ◦ in trocar orientation estimation, with an average processing time of 15 fps. This could serve as a foundation to improve robotic systems’ automation, integration, and efficiency of robotic systems for retinal surgery. The dataset created for this work is made publicly available.

关键词： Tracking Navigation Instruments Robot vision systems Surgery Manuals Retina

来源：评论

学校读者我要写书评

暂无评论

Single Frame Lidar-Camera Calibration Using Registration of 3D Planes

Single Frame Lidar-Camera Calibration Using Registration of ...

引用

IEEE International Conference on Robotic Computing (IRC)

作者： Ashutosh Singandhupe Hung Manh La Quang Phuc Ha Department of Computer Science and Engineering Advanced Robotics and Automation (ARA) Lab University of Nevada Reno NV USA School of Electrical and Data Engineering University of Technology Sydney Australia

ISBN: (纸本)9781665472616

This work focuses on finding the extrinsic parameters (rotation and translation) between Lidar and an RGB camera sensor. We use a planar checkerboard and place it inside the Field-of-View (FOV) of both sensors, where we extract the 3D plane information of the checkerboard acquired from the sensor’s data. The plane coefficients extracted from the sensor’s data are used to construct a well-structured set of 3D points. These 3D points are then ’aligned,’ which gives the relative transformation between the two sensors. We use our proposed Correntropy Similarity Matrix Iterative Closest Point (CoSMICP) Algorithm to estimate the relative transformation. This work uses a single frame of the point cloud data acquired from the Lidar sensor and a single frame from the calibrated camera data to perform this operation. From the camera image, we use the projection of the calibration target’s corner points to compute the 3D points, and along the process, we calculate the 3D plane equation using the corner points. We evaluate our approach on a simulated dataset with complex environment settings, making use of the freedom to assess under multiple configurations. Through the obtained results, we verify our method under various configurations.

关键词： Point cloud compression Three-dimensional displays Laser radar Transmission line matrix methods Robot vision systems Cameras Real-time systems

来源：评论

学校读者我要写书评

暂无评论

AirVO: An Illumination-Robust Point-Line Visual Odometry

AirVO: An Illumination-Robust Point-Line Visual Odometry

引用

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

作者： Kuan Xu Yuefan Hao Shenghai Yuan Chen Wang Lihua Xie School of Electrical and Electronic Engineering Nanyang Technological University Singapore Robot R&D Department Geekplus Corp. Beijing China Spatial AI & Robotics Lab at The Department of Computer Science & Engineering State University of New York at Buffalo Buffalo NY USA

This paper proposes an illumination-robust visual odometry (VO) system that incorporates both accelerated learning-based corner point algorithms and an extended line feature algorithm. To be robust to dynamic illumination, the proposed system employs the convolutional neural network (CNN) and graph neural network (GNN) to detect and match reliable and informative corner points. Then point feature matching results and the distribution of point and line features are utilized to match and triangulate lines. By accelerating CNN and GNN parts and optimizing the pipeline, the proposed system is able to run in real-time on low-power embedded platforms. The proposed VO was evaluated on several datasets with varying illumination conditions, and the results show that it outperforms other state-of-the-art VO systems in terms of accuracy and robustness. The open-source nature of the proposed system allows for easy implementation and customization by the research community, enabling further development and improvement of VO for various applications.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Semi-Supervised Learning for Intelligent Surveillance

Semi-Supervised Learning for Intelligent Surveillance

引用

IEEE Latin American robotics Symposium, LARS

作者： Guilherme Corrêa de Freitas Marcos R. O. A. Maximo Filipe A. N. Verri Computer Vision Department ALTAVE São José dos Campos SP Brazil Autonomous Computational Systems Lab (LAB-SCA) Computer Science Division Aeronautics Institute of Technology São José dos Campos SP Brazil

ISBN: (数字)9781665462808

ISBN: (纸本)9781665462815

Semi-supervised Learning (SSL) has shown promising improvements when used for the Object Detection (OD) task. Existing works detected some unique challenges associated with this setting and discussed how they differ from the classification paradigm. In this work, we revisit two of the first approaches used for OD, namely, Self-Training and Augmentation driven Consistency (STAC) and Unbiased Teacher. Our work was made in partnership with ALTAVE, a Brazilian company focused on intelligent surveillance. By using their in-house dataset we show that relative performance refinement follows the findings from the original authors and propose an improved augmentation method that boosts both approaches, obtaining an overall improvement of 4.4% in the mAP, while using no additional labels for training.

关键词： Training Terminology Surveillance Object detection Companies Machine learning Semisupervised learning

来源：评论

学校读者我要写书评

暂无评论

Analysis of Vegetation Indices in the Cotton Crop in South Asia region using UAV Imagery

Analysis of Vegetation Indices in the Cotton Crop in South A...

引用

International Conference on Emerging Technologies, ICET

作者： Noor us Sabah Muhammad Usama Zuhair Zafar Muhammad Shahzad Muhammad Moazam Fraz Karsten Berns School of Electrical Engineering and Computer Science National University of Sciences and Technology (NUST) Islamabad Pakistan Department of Aerospace and Geodesy Technical University of Munich (TUM) Munich Germany Department of Computer Science Robotics Research Lab Technische Universität Kaiserslautern (TUK) Kaiserslautern Germany

ISBN: (纸本)9781665459938

Cotton is one of the most economically significant crops in the South Asia region. Following the decline in cotton production in recent years, it is necessary to study and analyse cotton growth patterns by using modern precision agriculture techniques to enhance crop yield and quality. Multispectral imagery plays a vital role in estimating crop quantification by computing vegetation indices at different phenological stages of the crop. It facilitates in estimating plant health and impact on yield before the visible symptoms appear. In this study, four different vegetation indices maps are computed using two-band multispectral imagery acquired through Unmanned Aerial Vehicle (UAV). The UAV multispectral data is collected from two fields with different sowing dates using Sentera NDVI single sensor. The variation in the vegetation indices map generated is analysed by vegetation index thresholding and k-means clustering. By using Sentinel-2A imagery, an average value of each vegetation index for every phenological stage is validated. The correlation coefficient of each vegetation index conveys a positive progression trend and validates the accuracy of UAV data computed results. In conclusion, vegetation indices computed using multispectral UAV imagery conveys vital information regarding the progression of crop phenology and growth pattern.

关键词： Image resolution Asia Vegetation mapping Crops Autonomous aerial vehicles Market research Cotton

来源：评论

学校读者我要写书评

暂无评论

Cylindrical-Lens-Embedded Photonic Crystal Based on Self-Collimation 5

Cylindrical-Lens-Embedded Photonic Crystal Based on Self-Col...

引用

5th IEEE Research and Applications of Photonics in Defense Conference, RAPID 2022

作者： Xia, Chun Gutierrez, Jesus J. Kuebler, Stephen M. Rumpf, Raymond C. Touma, Jimmy University of Central Florida The College of Optics and Photonics Creol OrlandoFL32816 United States University of Texas at El Paso Em Lab Department of Electrical and Computer Engineering El PasoTX79968 United States Materials Science and Engineering University of Central Florida Creol The College of Optics and Photonics Chemistry Department OrlandoFL32816 United States Computational Science Program University of Texas at El Paso Em Lab Department of Electrical and Computer Engineering El PasoTX79968 United States Eglin Afb AFRL/Munitions Directorate Integrated Sensing and Processing Branch FL32542 United States

Independent optical power and phase control is experimentally demonstrated in a self-collimating PC based on rod-into-wall unit cell and fabricated in a low-refractive-index medium by multiphoton lithography. © 2... 详细信息

ISBN: (数字)9781665402231

ISBN: (纸本)9781665402231

关键词： Refractive index

来源：评论

学校读者我要写书评

暂无评论

Mechatronic Design for Multi Robots-Insect Swarms Interactions

Mechatronic Design for Multi Robots-Insect Swarms Interactio...

引用

2023 IEEE International Conference on Mechatronics (ICM)

作者： Fatemeh Rekabi-Bana Martin Stefanec Jiří Ulrich Erhan E. Keyvan Tomáš Rouček George Broughton Bilal Y. Gündeǧer Ömer Sahin Ali E. Turgut Erol Şahin Tomáš Krajník Thomas Schmickl Farshad Arvin Department of Computer Science Swarm & Computational Intelligence Laboratory (SwaCIL) Durham University UK Artificial Life Lab University of Graz Austria Faculty of Electrical Engineering Czech Technical University Czechia Middle East Technical University Türkiye

This paper presents the concept of a robotic system collaborating with a swarm of social insects inside their hive. This robot consists of a micro-and macro-manipulator and a tracking system. The micro-manipulator uses bio-mimetic agents to interact with an individual specimen. The macromanipulator positions and keeps the micro-manipulator’s base around the given individual while moving in the hive. This individual is tracked by a fiducial marker-based visual detection and localisation system, which also provides positions of the biomimetic agents. The base of the system was experimentally verified in a honeybee observation hive, where it flawlessly tracked the honeybee queen for several hours, gathering sufficient data to extract the behaviours of honeybee workers in the queen’s vicinity. These data were then used in simulation to verify if the micro-manipulator’s bio-mimetic agents could mimic some of the honeybee workers’ behaviours.

关键词： Visualization Mechatronics Tracking Databases Machine vision Insects Data models

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：