检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Wang, Yikai Dong, Yinpeng Sun, Fuchun Yang, Xiao State Key Lab on Intelligent Technology and Systems Department of Computer Science and Technology Tsinghua University China RealAI China

This work focuses on the 3D reconstruction of non-rigid objects based on monocular RGB video sequences. Concretely, we aim at building high-fidelity models for generic object categories and casually captured scenes. To this end, we do not assume known root poses of objects, and do not utilize category-specific templates or dense pose priors. The key idea of our method, Root Pose Decomposition (RPD), is to maintain a per-frame root pose transformation, meanwhile building a dense field with local transformations to rectify the root pose. The optimization of local transformations is performed by point registration to the canonical space. We also adapt RPD to multi-object scenarios with object occlusions and individual differences. As a result, RPD allows non-rigid 3D reconstruction for complicated scenarios containing objects with large deformations, complex motion patterns, occlusions, and scale diversities of different individuals. Such a pipeline potentially scales to diverse sets of objects in the wild. We experimentally show that RPD surpasses state-of-the-art methods on the challenging DAVIS, OVIS, and AMA datasets. We provide video results in https://***. Copyright © 2023, 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

Survey semantic Arabic language in deep learning based recommendation system

Survey semantic Arabic language in deep learning based recom...

引用

IEEE International Symposium on Innovations in intelligent systems and Applications (INISTA)

作者： Rasha Kadhem Souheyl Mallat Mounir Zrigu Department of Computer Science University of ALQadisiyah Iraq Research Laboratory in Algebra Numbers Theory and Intelligent Systems Monastir Tunisia Department of Computer Science Faculty of Sciences of Monastir University of Monastir

Due to problems, Arabic-speaking internet users have surged, although nothing is done on it. It is challenging to develop a repliable recognition system (RS) for cursive languages such as Arabic. Variations in text size, fonts, word semantics, user Arabic region, etc. complicate these issues. Deep learning models can model big datasets and handle them. Good features can be selected and learned consecutively by both CNNs and RNNs. In numerous studies, both of these neural networks have proven to be superior than their counterparts. This is true for text recognition, voice recognition, and several NLP tasks (NLP), but they are not qualified to deal with the semantics of the text, so we decided to find the best DL technique for semantic Arabic language from a lot of research to be a survey to other searchers. Our paper compared different algorithms and their accuracy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

UNCERTAINTY-PENALIZED DIRECT PREFERENCE OPTIMIZATION

arXiv

引用

arXiv 2024年

作者： Houliston, Sam Pace, Alizée Immer, Alexander Rätsch, Gunnar Department of Computer Science ETH Zurich Switzerland ETH AI Center Zurich Switzerland Max Planck Institute for Intelligent Systems Tübingen Germany

Aligning Large Language Models (LLMs) to human preferences in content, style, and presentation is challenging, in part because preferences are varied, context-dependent, and sometimes inherently ambiguous. While successful, Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) are prone to the issue of proxy reward overoptimization. Analysis of the DPO loss reveals a critical need for regularization for mislabeled or ambiguous preference pairs to avoid reward hacking. In this work, we develop a pessimistic framework for DPO by introducing preference uncertainty penalization schemes, inspired by offline reinforcement learning. The penalization serves as a correction to the loss which attenuates the loss gradient for uncertain samples. Evaluation of the methods is performed with GPT2 Medium on the Anthropic-HH dataset using a model ensemble to obtain uncertainty estimates, and shows improved overall performance compared to vanilla DPO, as well as better completions on prompts from high-uncertainty chosen/rejected *** Codes Learning and adaptive systems in artificial intelligence Copyright © 2024, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Multi-Player Potential Game Approach for Sensor Network Localization with Noisy Measurements 63

A Multi-Player Potential Game Approach for Sensor Network Lo...

引用

63rd IEEE Conference on Decision and Control, CDC 2024

作者： Xu, Gehui Chen, Guanpu Fidan, Baris Hong, Yiguang Qi, Hongsheng Parisini, Thomas Johansson, Karl H. Kth Royal Institute of Technology Division of Decision and Control Systems School of Electrical Engineering and Computer Science Stockholm100 44 Sweden University of Waterloo Department of Mechanical and Mechatronics Engineering WaterlooONN2L 3G1 Canada Tongji University Department of Control Science and Engineering shanghai201804 China Shanghai Research Institute for Intelligent Autonomous Systems Shanghai201210 China Key Laboratory of Systems and Control Academy of Mathematics and Systems Science Beijing China University of Chinese Academy of Sciences School of Mathematical Sciences Beijing China Aalborg University Department of Electronic Systems Denmark Imperial College London Department of Electrical and Electronic Engineering LondonSW7 2AZ United Kingdom University of Trieste Department of Engineering and Architecture Italy

ISBN: (纸本)9798350316339

Sensor network localization (SNL) is a challenging problem due to its inherent non-convexity and the effects of noise in inter-node ranging measurements and anchor node position. We formulate a non-convex SNL problem as a multiplayer non-convex potential game and investigate the existence and uniqueness of a Nash equilibrium (NE) in both the ideal setting without measurement noise and the practical setting with measurement noise. We first show that the NE exists and is unique in the noiseless case, and corresponds to the precise network localization. Then, we study the SNL for the case with errors affecting the anchor node position and the inter-node distance measurements. Specifically, we establish that in case these errors are sufficiently small, the NE exists and is unique. It is shown that the NE is an approximate solution to the SNL problem, and that the position errors can be quantified accordingly. Based on these findings, we apply the results to case studies involving only inter-node distance measurement errors and only anchor position information inaccuracies. © 2024 IEEE.

关键词： Nash equilibrium

来源：评论

学校读者我要写书评

暂无评论

How robust accuracy suffers from certified training with convex relaxations

arXiv

引用

arXiv 2023年

作者： De Bartolomeis, Piersilvio Clarysse, Jacob Sanyal, Amartya Yang, Fanny Department of Computer Science ETH Zürich Switzerland Max Planck Institute for Intelligent Systems Tübingen Germany

Adversarial attacks pose significant threats to deploying state-of-the-art classifiers in safety-critical applications. Two classes of methods have emerged to address this issue: empirical defences and certified defences. Although certified defences come with robustness guarantees, empirical defences such as adversarial training enjoy much higher popularity among practitioners. In this paper, we systematically compare the standard and robust error of these two robust training paradigms across multiple computer vision tasks. We show that in most tasks and for both ∞-ball and 2-ball threat models, certified training with convex relaxations suffers from worse standard and robust error than adversarial training. We further explore how the error gap between certified and adversarial training depends on the threat model and the data distribution. In particular, besides the perturbation budget, we identify as important factors the shape of the perturbation set and the implicit margin of the data distribution. We support our arguments with extensive ablations on both synthetic and image datasets. © 2023, CC BY.

关键词： Errors

来源：评论

学校读者我要写书评

暂无评论

Human-Factor focused Application of STPA to Remotely driven Vehicles

Human-Factor focused Application of STPA to Remotely driven ...

引用

Public Safety Technology (WFPST), IEEE World Forum on

作者： Ole Hans Krystian Radlak Jürgen Adamy Department of Control Methods and Intelligent Systems Technical University of Darmstadt Düsseldorf Germany Institute of Computer Science Warsaw University of Technology Warsaw Poland

ISBN: (数字)9798350329148

ISBN: (纸本)9798350329155

This paper presents an analysis of the safety aspects of a Remote Driving System (RDS) using the System-Theoretic Process Analysis (STPA) methodology. The focus is on the human Remote Driver (RD) within the RDS, considering losses and system-level hazards associated with the human-in-the-loop scenario which focuses on the RD actions related to the change in Steering Wheel Angle (SWA). The STPA methodology is applied to identify Unsafe Control Actions (UCA) related to the change in SWA command, resulting in the identification of numerous Causal Factors (CF) that may lead to hazardous scenarios. Several UCAs and CFs are identified that provide information on potential hazards and risks associated with remote driving technology. Based on a selected UCA where the RD does not change the SWA when the road network requires lateral movement, 25 potential loss scenarios were identified and presented in this paper. The results of this work can contribute to the understanding of potential risks and the development of safety mitigation measures in the context of RDS which need to be validated in further research.

关键词： Roads Prevention and mitigation Wheels Hazards Public security Human in the loop Behavioral sciences

来源：评论

学校读者我要写书评

暂无评论

Automated Parts Segmentation of Sperm via a Contrastive Learning-Based Part Matching Network

引用

IEEE Transactions on Automation science and Engineering 2025年 22卷 15314-15325页

作者： Chen, Wenyuan Song, Haocong Shan, Guanqiao Dai, Changsheng Liu, Hang Jiang, Aojun Sun, Chen Ru, Changhai Librach, Clifford Zhang, Zhuoran Sun, Yu University of Toronto Department of Computer Science TorontoONM5S 2E4 Canada University of Toronto Department of Mechanical and Industrial Engineering TorontoONM5S 3G8 Canada Dalian University of Technology Institute of Robotics and Intelligent Systems Liaoning Dalian China Suzhou University of Science and Technology School of Electronic and Information Engineering Suzhou215009 China CReATe Fertility Centre TorontoONM5G 1N8 Canada

Sperm morphology measurement is vital for diagnosing male infertility, which involves quantification of multiple subcellular parts for each sperm. Instance-aware part segmentation networks have been introduced to address this task by automatically identifying individual sperm and segmenting their subcellular parts. However, major limitations of state-of-the-art instance-aware part segmentation networks include: 1) they are time-consuming and computational expensive due to sequential processing and multi-stage frameworks;2) they perform poorly for densely packed sperm that overlap or cross over one another. To overcome these challenges, this paper proposes 1) integrating instance identification and subcellular part segmentation within a single-stage framework to save inference time and memory usage;2) dividing a sperm target into simpler components (head and tail) to improve prediction accuracy, followed by a contrastive learning-based matching method to pair the head and tail. Experimental results on our clinically collected human sperm dataset demonstrated that the proposed network not only outperformed state-of-the-art CP-Net (by 3.5% APpvol) but also achieved realtime inference (48.0 frames per second), effectively meeting the clinical requirements for automated parts segmentation of sperm. © 2004-2012 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Human Fall Detection Using Transfer Learning-Based 3D CNN 1st

Human Fall Detection Using Transfer Learning-Based 3D CNN

引用

1st International Conference on Computational Technologies and Electronics, ICCTE 2023

作者： Alam, Ekram Sufian, Abu Dutta, Paramartha Leo, Marco Department of Computer Science Gour Mahavidyalaya Old Malda India Department of Computer Science University of Gour Banga English Bazar India Department of Computer and System Sciences Visva-Bharati University Santiniketan India National Research Council of Italy Institute of Applied Sciences and Intelligent Systems Lecce Italy

ISBN: (纸本)9783031819346

Unintentional or accidental falls are one of the significant health issues in senior persons. The population of senior persons is increasing steadily. So, there is a need for an automated fall detection monitoring system. This paper introduces a vision-based fall detection system using a pre-trained 3D CNN. Unlike 2D CNN, 3D CNN extracts not only spatial but also temporal features. Proposed model leverages the original learned weights of a 3D CNN model pre-trained on the Sports1M dataset to extract the spatio-temporal features. Only the SVM classifier was trained, which saves the time required to train the 3D CNN. Stratified shuffle five split cross-validation has been used to split the dataset into training and testing data. Extracted features from the proposed 3D CNN model were fed to an SVM classifier to classify the activity as fall or ADL. Two datasets, GMDCSA and CAUCAFall, were utilized to conduct the experiment. The source code for this work can be accessed via the following link: https://***/ekramalam/HFD_3DCNN. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Fall detection

来源：评论

学校读者我要写书评

暂无评论

Source Camera Device Identification from Videos

引用

SN computer science 2022年第4期3卷 316页

作者： Bennabhaktula, Guru Swaroop Timmerman, Derrick Alegre, Enrique Azzopardi, George Information Systems Group at the Bernoulli Institute for Mathematics Computer Science and Artificial Intelligence University of Groningen Groningen Netherlands Group for Vision and Intelligent Systems at the Department of Electrical Systems and Automation University of León León Spain Researcher at INCIBE (Spanish National Cybersecurity Institute) León Spain

Source camera identification is an important and challenging problem in digital image forensics. The clues of the device used to capture the digital media are very useful for Law Enforcement Agencies (LEAs), especially to help them collect more intelligence in digital forensics. In our work, we focus on identifying the source camera device based on digital videos using deep learning methods. In particular, we evaluate deep learning models with increasing levels of complexity for source camera identification and show that with such sophistication the scene-suppression techniques do not aid in model performance. In addition, we mention several common machine learning strategies that are counter-productive in achieving a high accuracy for camera identification. We conduct systematic experiments using 28 devices from the VISION data set and evaluate the model performance on various video scenarios—flat (i.e., homogeneous), indoor, and outdoor and evaluate the impact on classification accuracy when the videos are shared via social media platforms such as YouTube and WhatsApp. Unlike traditional PRNU-noise (Photo Response Non-Uniform)-based methods which require flat frames to estimate camera reference pattern noise, the proposed method has no such constraint and we achieve an accuracy of 72.75 ± 1.1 % on the benchmark VISION data set. Furthermore, we also achieve state-of-the-art accuracy of 71.75 % on the QUFVD data set in identifying 20 camera devices. These two results are the best ever reported on the VISION and QUFVD data sets. Finally, we demonstrate the runtime efficiency of the proposed approach and its advantages to LEAs. © 2022, The Author(s).

关键词： Constrained ConvNet Digital video forensics Sensor pattern noise Source camera identification

来源：评论

学校读者我要写书评

暂无评论

GSLB: The Graph Structure Learning Benchmark 37

GSLB: The Graph Structure Learning Benchmark

引用

37th Conference on Neural Information Processing systems, NeurIPS 2023

作者： Li, Zhixun Wang, Liang Sun, Xin Luo, Yifan Zhu, Yanqiao Chen, Dingshuo Luo, Yingtao Zhou, Xiangxin Liu, Qiang Wu, Shu Yu, Jeffrey Xu Department of Systems Engineering and Engineering Management The Chinese University of Hong Kong Hong Kong Center for Research on Intelligent Perception and Computing State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Chinese Academy of Sciences China School of Artificial Intelligence University of Chinese Academy of Sciences China Department of Automation University of Science and Technology of China China School of Cyberspace Security Beijing University of Posts and Telecommunications China Department of Computer Science University of California Los Angeles United States Heinz College of Information Systems and Public Policy Machine Learning Department School of Computer Science Carnegie Mellon University United States

ISBN: (纸本)9781713899921

Graph Structure Learning (GSL) has recently garnered considerable attention due to its ability to optimize both the parameters of Graph Neural Networks (GNNs) and the computation graph structure simultaneously. Despite the proliferation of GSL methods developed in recent years, there is no standard experimental setting or fair comparison for performance evaluation, which creates a great obstacle to understanding the progress in this field. To fill this gap, we systematically analyze the performance of GSL in different scenarios and develop a comprehensive Graph Structure Learning Benchmark (GSLB) curated from 20 diverse graph datasets and 16 distinct GSL algorithms. Specifically, GSLB systematically investigates the characteristics of GSL in terms of three dimensions: effectiveness, robustness, and complexity. We comprehensively evaluate state-of-the-art GSL algorithms in node- and graph-level tasks, and analyze their performance in robust learning and model complexity. Further, to facilitate reproducible research, we have developed an easy-to-use library for training, evaluating, and visualizing different GSL methods. Empirical results of our extensive experiments demonstrate the ability of GSL and reveal its potential benefits on various downstream tasks, offering insights and opportunities for future research. The code of GSLB is available at: https://***/GSL-Benchmark/GSLB. © 2023 Neural information processing systems foundation. All rights reserved.

关键词： Graphic methods

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：