检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Zhang, Sixiao Chen, Hongxu Yang, Haoran Sun, Xiangguo Yu, Philip S. Xu, Guandong Data Science and Machine Intelligence Lab Faculty of Engineering and Information Technology University of Technology Sydney SydneyNSW2007 Australia The Chinese University of Hong Kong Hong Kong Department of Computer Science University of Illinois at Chicago ChicagoIL60637 United States

Recently, transformers have shown promising performance in learning graph representations. However, there are still some challenges when applying transformers to real-world scenarios due to the fact that deep transformers are hard to train from scratch and the quadratic memory consumption w.r.t. the number of nodes. In this paper, we propose Graph Masked Autoencoders (GMAEs), a self-supervised transformer-based model for learning graph representations. To address the above two challenges, we adopt the masking mechanism and the asymmetric encoder-decoder design. Specifically, GMAE takes partially masked graphs as input, and reconstructs the features of the masked nodes. The encoder and decoder are asymmetric, where the encoder is a deep transformer and the decoder is a shallow transformer. The masking mechanism and the asymmetric design make GMAE a memory-efficient model compared with conventional transformers. We show that, when serving as a conventional self-supervised graph representation model, GMAE achieves state-of-the-art performance on both the graph classification task and the node classification task under common downstream evaluation protocols. We also show that, compared with training in an end-to-end manner from scratch, we can achieve comparable performance after pre-training and fine-tuning using GMAE while simplifying the training process. Copyright © 2022, The Authors. All rights reserved.

关键词： Data mining

来源：评论

学校读者我要写书评

暂无评论

NavCoT: Boosting LLM-Based vision-and-Language Navigation via Learning Disentangled Reasoning

引用

IEEE Transactions on Pattern Analysis and machine intelligence 2025年第7期47卷 5945-5957页

作者： Bingqian Lin Yunshuang Nie Ziming Wei Jiaqi Chen Shikui Ma Jianhua Han Hang Xu Xiaojun Chang Xiaodan Liang Shenzhen Campus Sun Yat-sen University Shenzhen China Shanghai Jiao Tong University Shanghai China University of Hong Kong Pok Fu Lam Hong Kong Dataa Robotics Company Beijing China Huawei Noah's Ark Lab Shanghai China School of Information Science and Technology University of Science and Technology of China Hefei China Department of Computer Vision Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) Abu Dhabi UAE Peng Cheng Laboratory Shenzhen China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China

vision-and-Language Navigation (VLN), as a crucial research problem of Embodied AI, requires an embodied agent to navigate through complex 3D environments following natural language instructions. Recent research has highlighted the promising capacity of large language models (LLMs) in VLN by improving navigational reasoning accuracy and interpretability. However, their predominant use in an offline manner usually suffers from substantial domain gap between the VLN task and the LLM training corpus. This paper proposes a novel strategy called Nav igational C hain- o f- T hought (NavCoT), where we fulfill parameter-efficient in-domain training to enable self-guided navigational decision, leading to a significant mitigation of the domain gap in a cost-effective manner. Specifically, at each timestep, the LLM is prompted to forecast the navigational chain-of-thought by: 1) acting as a world model to imagine the next observation according to the instruction, 2) selecting the candidate observation that best aligns with the imagination, and 3) determining the action based on the reasoning from the prior steps. In this way, the action prediction can be effectively simplified benefiting from the disentangled reasoning. Through constructing formalized labels for training, the LLM can learn to generate desired and reasonable chain-of-thought outputs for improving the action decision. Experimental results across various training settings and popular VLN benchmarks (e.g., Room-to-Room (R2R), Room-across-Room (RxR), Room-for-Room (R4R)) show the significant superiority of NavCoT over the direct action prediction variants. Through simple parameter-efficient finetuning, our NavCoT outperforms a recent GPT4-based approach with $\sim$ 7% relative improvement on the R2R dataset. We believe that NavCoT will help unlock more task-adaptive and scalable LLM-based embodied agents, which are helpful for developing real-world robotics applications.

关键词： Navigation Cognition Training Scalability Glass Planning Large language models History Electronic mail Visualization

来源：评论

学校读者我要写书评

暂无评论

Neural architecture dilation for adversarial robustness

arXiv

引用

arXiv 2021年

作者： Li, Yanxi Yang, Zhaohui Wang, Yunhe Xu, Chang School of Computer Science University of Sydney Australia Noah's Ark Lab Huawei Technologies China Department of Machine Intelligence Peking University China

With the tremendous advances in the architecture and scale of convolutional neural networks (CNNs) over the past few decades, they can easily reach or even exceed the performance of humans in certain tasks. However, a recently discovered shortcoming of CNNs is that they are vulnerable to adversarial attacks. Although the adversarial robustness of CNNs can be improved by adversarial training, there is a trade-off between standard accuracy and adversarial robustness. From the neural architecture perspective, this paper aims to improve the adversarial robustness of the backbone CNNs that have a satisfactory accuracy. Under a minimal computational overhead, the introduction of a dilation architecture is expected to be friendly with the standard performance of the backbone CNN while pursuing adversarial robustness. Theoretical analyses on the standard and adversarial error bounds naturally motivate the proposed neural architecture dilation algorithm. Experimental results on real-world datasets and benchmark neural networks demonstrate the effectiveness of the proposed algorithm to balance the accuracy and adversarial robustness. Copyright © 2021, The Authors. All rights reserved.

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

Deep knn for medical image classification 23rd

Deep knn for medical image classification

引用

23rd International Conference on Medical Image Computing and computer-Assisted Intervention, MICCAI 2020

作者： Zhuang, Jiaxin Cai, Jiabin Wang, Ruixuan Zhang, Jianguo Zheng, Wei-Shi School of Data and Computer Science Sun Yat-sen University Guangzhou China Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China Key Laboratory of Machine Intelligence and Advanced Computing MOE Guangzhou China Pazhou Lab Guangzhou China

ISBN: (纸本)9783030597092

Human-level diagnostic performance from intelligent systems often depends on large set of training data. However, the amount of available data for model training may be limited for part of diseases, which would cause the widely adopted deep learning models not generalizing well. One alternative simple approach to small class prediction is the traditional k-nearest neighbor (kNN). However, due to the non-parametric characteristics of kNN, it is difficult to combine the kNN classification into the learning of feature extractor. This paper proposes an end-to-end learning strategy to unify the kNN classification and the feature extraction procedure. The basic idea is to enforce that each training sample and its K nearest neighbors belong to the same class during learning the feature extractor. Experiments on multiple small-class and class-imbalanced medical image datasets showed that the proposed deep kNN outperforms both kNN and other strong classifiers. © Springer Nature Switzerland AG 2020.

关键词： Nearest neighbor search

来源：评论

学校读者我要写书评

暂无评论

Fetal Re-Identification in Multiple Pregnancy Ultrasound Images Using Deep Learning

Fetal Re-Identification in Multiple Pregnancy Ultrasound Ima...

引用

Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

作者： Elisabeth Gabler Michael Nissen Thomas R. Altstidl Adriana Titzmann Kai Packhäuser Andreas Maier Peter A. Fasching Bjoern M. Eskofier Heike Leutheuser Department Artificial Intelligence in Biomedical Engineering Machine Learning and Data Analytics (MaD) Lab Friedrich-Alexander-Universität Erlangen-Nurnberg (FAU) Erlangen Germany Department of Gynecology and Obstetrics Erlangen University Hospital Friedrich-Alexander-Universität Erlangen-Nurnberg (FAU) Erlangen Germany Department of Computer Science Pattern Recognition Lab Friedrich-Alexander-Universität Erlangen-Nurnberg (FAU) Erlangen Germany

Ultrasound examinations during pregnancy can detect abnormal fetal development, which is a leading cause of perinatal mortality. In multiple pregnancies, the position of the fetuses may change between examinations. The individual fetus cannot be clearly identified. Fetal re-identification may improve diagnostic capabilities by tracing individual fetal changes. This work evaluates the feasibility of fetal re-identification on FETAL_PLANES_DB, a publicly available dataset of singleton pregnancy ultrasound images. Five dataset subsets with 6,491 images from 1,088 pregnant women and two re-identification frameworks (Torchreid, FastReID) are evaluated. FastReID achieves a mean average precision of 68.77% (68.42%) and mean precision at rank 10 score of 89.60% (95.55%) when trained on images showing the fetal brain (abdomen). Visualization with gradient-weighted class activation mapping shows that the classifiers appear to rely on anatomical features. We conclude that fetal re-identification in ultrasound images may be feasible. However, more work on additional datasets, including images from multiple pregnancies and several subsequent examinations, is required to ensure and investigate performance stability and *** relevance— To date, fetuses in multiple pregnancies cannot be distinguished between ultrasound examinations. This work provides the first evidence for feasibility of fetal re-identification in pregnancy ultrasound images. This may improve diagnostic capabilities in clinical practice in the future, such as longitudinal analysis of fetal changes or abnormalities.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Parallel bi-objective evolutionary algorithms for scalable feature subset selection via migration strategy under Spark

arXiv

引用

arXiv 2022年

作者： Vivek, Yelleti Ravi, Vadlamani Krishna, P. Radha Analytics Lab Centre for Artificial Intelligence and Machine Learning Institute for Development and Research in Banking Technology Castle Hills Road #1 Masab Tank Hyderabad500076 India Department of Computer Science and Engineering National Institute of Technology Warangal506004 India

Feature subset selection (FSS) for classification is inherently a bi-objective optimization problem, where the task is to obtain a feature subset which yields the maximum possible area under the receiver operator characteristic curve (AUC) with minimum cardinality of the feature subset. In today's world, humungous amount of data is generated in all activities of humans. To mine such voluminous data, which is often high-dimensional, there is a need to develop parallel and scalable frameworks. In the first-of-its-kind study, we propose and develop an iterative MapReduce based framework for bi-objective evolutionary algorithms (EAs) based wrappers under Apache spark with the migration strategy. In order to accomplish this, we parallelized the non-dominated sorting based algorithms namely non dominated sorting algorithm (NSGA-II), and non-dominated sorting particle swarm optimization (NSPSO), also the decomposition based algorithm, namely the multi objective evolutionary algorithm based on decomposition (MOEA/D), and named them P-NSGA-II-IS, P-NSPSO-IS, P-MOEA/D-IS, respectively. We proposed a modified MOEA/D by incorporating the non-dominated sorting principle while paralleizing it. Throughout the study, AUC is computed by logistic regression (LR). We test the effectiveness of the proposed methodology on various datasets. It is noteworthy that the P-NSGA-II turns out to be statistically significant by being in the top 2 positions on most datasets. We also reported the empirical attainment plots, speedup analysis, and mean AUC obtained by the most repeated feature subset and the least cardinal feature subset with the highest AUC, and diversity analysis using *** Codes 68T09 © 2022, CC BY-NC-ND.

关键词： Particle swarm optimization (PSO)

来源：评论

学校读者我要写书评

暂无评论

Feature Subset Selection for Big Data Via Parallel Chaotic Binary Differential Evolution Island Model Under Apache Spark

SSRN

引用

SSRN 2022年

作者： Yelleti, Vivek Vadlamani, Ravi Radha Krishna, P. Analytics Lab Centre for Artificial Intelligence Machine Learning Institute for Development and Research in Banking Technology Castle Hills Road #1 Masab Tank Hyderabad500076 India Department of Computer Science and Engineering National Institute of Technology Warangal506004 India

Feature subset selection (FSS) using a wrapper approach is essentially a combinatorial optimization problem having two objective functions viz., (i) the cardinality of the selected feature subset, and (ii) the area under the ROC curve (AUC) of a classifier built on this subset. They should be minimized and maximized respectively. In this paper, we propose a novel multiplicative single objective function involving cardinality and AUC for FSS in high-dimensional and large-scale datasets. Since the randomness involved in the Binary Differential Evolution (BDE) may yield less diverse solutions and get the algorithm entrapped in local minima, we embed Logistic and Tent Chaotic maps into the BDE, called Chaotic Binary Differential Evolution (CBDE). We (i) develop an adaptive version of CBDE based on SHADE (called CSHADE) and (ii) propose a scalable island (IS) based parallelization approach, where the data is divided into multiple partitions/islands thereby enabling the solutions evolve individually and get combined eventually in a migration strategy/policy. The results empirically show that the proposed parallel Chaotic BDE (P-CBDE-IS) and its corresponding adaptive version can find better quality feature subsets than the Parallel BDE (P-BDE-IS). Overall, P-CBDE-IS performed best in most of the cases. Logistic Regression (LR) is used as a classifier owing to its simplicity and power. The speedup attained by the proposed parallel approach signifies its importance. © 2022, The Authors. All rights reserved.

关键词： Large dataset

来源：评论

学校读者我要写书评

暂无评论

Real-Time Data Collection of Wsn Based on Trajectory Optimization of Laser-Charged Uav

SSRN

引用

SSRN 2023年

作者： Luo, Chuanwen Hong, Yi Hou, Yunan Zhu, Yuqing Li, Deying Yu, Dongxiao Wu, Weili School of Information Science and Technology Beijing Forestry University Beijing100083 China Engineering Research Center for Forestry-oriented Intelligent Information Processing of National Forestry and Grassland Administration Beijing100083 China Research Center of Applied Mathematics and Machine Intelligence Zhejiang Lab Zhejiang Hangzhou311121 China Department of Computer Science California State University Los AngelesCA90032 United States School of Information Renmin University of China Beijing100872 China School of Computer Science and Technology Institute of Intelligent Computing Shandong University Qingdao266510 China Department of Computer Science The University of Texas at Dallas RichardsonTX75080 United States

Unmanned Aerial Vehicle (UAV) can be used as wireless aerial mobile base station for collecting data from sensors in UAV-based Wireless Sensor Networks (WSNs), which is crucial for providing seamless services and improving the performance in the next generation wireless networks. However, since the UAV are powered by batteries with limited energy capacity, the UAV can not complete data collection tasks of all sensors without energy replenishment when a large number of sensors are deployed over large monitoring areas. To overcome this problem, this paper study the Real-time Data Collection with Laser-charging UAV (RDCL) problem, where the UAV is used to collect data from a given WSN and replenished energy by Laser Beam Directors (LBDs). The problem aims at optimizing the flight trajectory of UAV for ensuring the real-time performance of data, which is to design the detailed hovering and traveling plans of UAV for collecting data from sensors and charged by LBDs such that all sensory data of WSN are collected by UAV and transported to the base *** prove that the RDCL problem is NP-hard. To solve the problem, we first study two sub-problems, the Trajectory Optimization of UAV for Data Collection (TODC) problem and the Charging Trajectory Optimization of UAV (CTO) problem, whose objectives are to find the optimal flight plans of UAV in the data collection areas and charging areas, respectively. Then we propose an approximation algorithm to solve each of them with the constant factor. Afterwards, we propose an approximation algorithm to solve the RDCL problem based on TODC and CTO problems. Finally, the proposed algorithm is verified by extensive simulations. © 2023, The Authors. All rights reserved.

关键词： Wireless sensor networks

来源：评论

学校读者我要写书评

暂无评论

Maximal η-clique maintenance over uncertain graph streams

引用

Information sciences 2025年 717卷

作者： Ma, Ziyi Wang, Liqing Yang, Jianye Zhou, Xu Li, Kenli Gao, Cuiyun School of Artificial Intelligence Hebei University of Technology Tianjin300401 China Chinacoal Beijing Coal Mining Machinery Co. Ltd. Beijing102400 China Guangxi Key Laboratory of Machine Vision and Intelligent Control Wuzhou University Wuzhou543002 China Department of New Networks PengCheng Laboratory Shenzhen518055 China Cyberspace Institute of Advanced Technology Guangzhou University Guangzhou510006 China College of Computer Science and Electronic Engineering Hunan University Changsha410082 China Shenzhen518055 China

Maximal clique enumeration is a critical task for analyzing graph data and has a wide range of applications, such as community detection, protein complex identification, and group recommendation. Although many efficient algorithms have been proposed, most of them are designed for static or deterministic graphs. However, in many real-life applications, graphs are often highly dynamic and nondeterministic. In this paper, we study the problem of maximal η-clique maintenance (shorted as MCM) over uncertain graph streams. Given an uncertain graph G and a probability threshold η, MCM reports all incremental η-cliques in G for each update edge of G. To the best of our knowledge, we are the first to systematically study this problem. We first develop a local pivot-based enumeration technique that utilizes the pivot priority of candidates to achieve effective pruning during the maximal η-clique enumeration procedure. Based on the technique, we propose a competitive baseline approach LPMCM, which finds newly generated maximal η-cliques and internal maximal η-cliques. Although LPMCM avoids repeated enumeration of the same maximal η-cliques, it still requires expensive costs in terms of η-clique maximality determination. To further enhance maintenance efficiency, we introduce an approach HLPMCM using the hash-based maximality determination technique and an improved internal maximal η-clique enumeration technique. Extensive experiments show that HLPMCM can achieve 1.1x-19.0x speedup over the competitors and consumes slightly more memory. © 2025 Elsevier Inc.

关键词： Maintenance

来源：评论

学校读者我要写书评

暂无评论

On the analysis of HEVC Intra Prediction Mode Decision Variants

引用

Procedia computer science 2020年 171卷 1887-1897页

作者： Preethi S. Nair Madhu S. Nair Department of Computer Science University of Kerala Kariavattom Thiruvananthapuram-695581 Kerala India Artificial Intelligence & Computer Vision Lab Department of Computer Science Cochin University of Science and Technology Kochi -682022 Kerala India

This paper presents a comparative analysis of some major High Efficiency Video Coding (HEVC) intra prediction mode decision methods reported in the literature. Intra coding in HEVC is based on spatial sample prediction and can be described as a collection of many elements including quad tree partitioning, 35 intra prediction modes, reference sample smoothing, boundary sample filtering process, intra mode coding and so on. Each of these elements results in an improved coding efficiency in HEVC intra coding. In this review, we provide an extensive and up-to-date discussion of the research proposals regarding a variety of HEVC intra prediction mode decision strategies. Moreover, some most significant research contributions on fast intra mode decision methods in HEVC to improvise the encoder efficiency are reviewed in detail along with a comparative performance analysis of the reviewed proposals in terms of some standard evaluation metrics such as bit rate, PSNR and time savings. In addition, this paper point out some research directions towards further developments in the intra prediction for the betterment of the encoder efficacy.

关键词： HEVC Intra prediction Mode decision RMD RDO

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：