检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Liu, Liu Liu, Xuanqing Hsieh, Cho-Jui Tao, Dacheng Institute of Artificial Intelligence State Key Lab of Software Development Environment Beihang University China University of California Los Angeles United States School of Computer Science Sydney AI Centre Faculty of Engineering the University of Sydney SydneyNSW2008 Australia

Trust-region (TR) and adaptive regularization using cubics (ARC) have proven to have some very appealing theoretical properties for non-convex optimization by concurrently computing function value, gradient, and Hessian matrix to obtain the next search direction and the adjusted parameters. Although stochastic approximations help largely reduce the computational cost, it is challenging to theoretically guarantee the convergence rate. In this paper, we explore a family of stochastic TR and ARC methods that can simultaneously provide inexact computations of the Hessian matrix, gradient, and function values. Our algorithms require much fewer propagations overhead per iteration than TR and ARC. We prove that the iteration complexity to achieve ǫ-approximate second-order optimality is of the same order as the exact computations demonstrated in previous studies. Additionally, the mild conditions on inexactness can be met by leveraging a random sampling technology in the finite-sum minimization problem. Numerical experiments with a non-convex problem support these findings and demonstrate that, with the same or a similar number of iterations, our algorithms require less computational overhead per iteration than current second-order methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

Spatial-temporal Consistency based Crowd Emotion Recognition

Spatial-temporal Consistency based Crowd Emotion Recognition

引用

International Conference on computer Supported Cooperative Work in Design

作者： Chuanmiao Zhao Guijuan Zhang Lei Lv Hong Liu Dianjie Lu School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Jinan China State Key Laboratory of High-end Server & Storage Technology Jinan China

In crowd evacuation scenarios, it is an effective way to ensure evacuation safety by recognizing real crowd emotions and then taking measures such as emotional infection to reduce crowd panic. However, factors such as distance, exposure, angle, and occlusion can lead to incomplete face expression information collection, thus failing to accurately identify crowd emotions. Therefore, it is still a very challenging problem to recognize individual emotions in the case of incomplete face expression information collection, and thus accurately identify crowd emotions during evacuation. To solve this problem, we propose a spatial-temporal consistency-based crowd emotion recognition method to accurately identify the real emotions of the crowd. First, for video frames that can capture the complete facial expression information, we use the residual network to accurately identify the individual emotion values in each frame. For video frames that cannot capture the complete facial expression information, we propose an individual emotion calculation model based on spatial-temporal consistency to calculate the individual emotion values in each frame. Second, we define the crowd panic level and obtain the real crowd emotion by calculation. Finally, we implement an end-to-end crowd panic emotion recognition system to verify our method. The experimental results show that the method can accurately calculate the crowd panic level, which is important for guiding crowd evacuation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Cross-Domain Animal Pose Estimation with Skeleton Anomaly-Aware Learning

引用

IEEE Transactions on Circuits and Systems for Video Technology 2025年

作者： Han, Le Chen, Kaixuan Zhao, Lei Jiang, Yangbo Wang, Pengfei Zheng, Nenggan Zhejiang Hangzhou310007 China Zhejiang University College of Computer Science and Technology Zhejiang Hangzhou310007 China Zhejiang University State Key Laboratory of Blockchain and Data Security Zhejiang Hangzhou310007 China Institute of Blockchain and Data Security. China Zhejiang University School of Software Technology Ningbo China Zhejiang University State Key Lab of Brain-Machine Intelligence Hangzhou310007 China Zhejiang Provincial Government ZJU Collaborative Innovation Center for Artificial Intelligence by MOE Hangzhou310007 China Bengbu University School of Computer and Information Engineering Bengbu233030 China

Animal pose estimation is often constrained by the scarcity of annotations and the diversity of scenarios and species. The pseudo-label generation based unsupervised domain adaptation paradigm, which discriminates the predicted keypoints of unlabeled data based on the skeleton position consistency, has demonstrated effectiveness for such problems. However, existing methods generate pseudo-labels with massive false positives, because they cannot effectively distinguish sample pairs with the same errors. In this study, we propose a cross-domain animal pose estimation model from a novel perspective of skeleton anomaly learning. We construct a graph contrastive learning mechanism to acquire the skeleton anomaly-aware knowledge, which enables the generation of accurate pseudo-labels for target domain and imposes graph constraint on unlabeled data. And a skeleton anomaly-feedback based domain adaptation framework is designed to facilitate implicit alignment of object-specific features and joint training of cross-domain. Besides, we propose a novel rat pose dataset named UDARP-9.4K to address the gap of small-sized animal pose datasets encompassing diverse experimental scenarios. The related datasets are reviewed and evaluated in detail. Extensive experiments are conducted on UDARP-9.4K and two public datasets to demonstrate the superiority of the proposed model in cross-scenarios and cross-species animal pose estimation tasks. Further analysis reveals the effectiveness of the proposed model for skeleton structure feature learning. The UDARP-9.4K dataset is available here. © 1991-2012 IEEE.

关键词： Invertebrates

来源：评论

学校读者我要写书评

暂无评论

An Epidemic Model Based on Intra- and Inter-group Interactions

An Epidemic Model Based on Intra- and Inter-group Interactio...

引用

International Conference on computer Supported Cooperative Work in Design

作者： Wencong Geng Guijuan Zhang Dianjie Lu School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Jinan China State Key Laboratory of High-end Server & Storage Technology Jinan China

The global spread of COVID-19 causes great losses to human society. Accurate calculation of the scale of epidemic spread is of great significance for the implementation of corresponding epidemic prevention measures. However, the existing method ignores the group formed by social relations of the population, which reduces the accuracy of the epidemic spread number calculation. In this paper, we propose an epidemic model based on intra- and inter-group interactions. Firstly, we construct a dual network model of epidemic spread based on intra- and inter-group interactions. The network describes how epidemics spread intra- and inter-group. To capture the intergroup influences, we construct a model for social mobility to calculate the inter-group spread rate. Secondly, we propose a computational model for the epidemic spread. We calculate the infection probability of groups in the upper layer network by using a continuous-time Markov chain (CTMC). We describe a dynamic evolution of the intra-group infection in the underlying network based on the mean field equation. And the number of infections in the population is calculated by integrating intra- and inter-group effects. Finally, we implement an epidemic spread simulation system to visualize the spread process. The experimental results show that the model can analyze the epidemic spread process more accurately.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Singular Value Decomposition Based Pedestrian Trajectory Prediction

Singular Value Decomposition Based Pedestrian Trajectory Pre...

引用

International Conference on computer Supported Cooperative Work in Design

作者： Heng Wang Guijuan Zhang Hong Liu Dianjie Lu School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Jinan China State Key Laboratory of High-end Server & Storage Technology Jinan China

Numerous cameras deployed in venues can collect video data, which can be analyzed to help evacuate people in emergency situations. Most surveillance videos can provide data support for pedestrian trajectory prediction. However, data-driven prediction method do not consider the impact of personalization on pedestrian trajectories, which results in ignoring individual personalization. How to accurately analyze the differences among pedestrians and yet accurately predict pedestrian trajectories is a challenging problem. To solve the above problems, we propose an singular value decomposition (SVD) based pedestrian trajectory prediction method, which introduces matrix decomposition to pedestrian trajectory prediction for the first time. Thus, This method analyzes the impact of pedestrian personalization on motion trajectories by mining the interaction patterns between pedestrians and environment. Firstly, we propose an information collection method based on environmental semantics, which can extract scene information from historical video data to construct an environmental information matrix. Secondly, we propose an SVD-based individual environmental feature preference method, which uses singular value decomposition methods to mine the data and analyze personalized pedestrian motion patterns to construct a personalized individual preference matrix. Finally, we built an personalized trajectory prediction method to predict pedestrian movement trajectories. The experimental results show that the method can not only analyze the effect of personalization on pedestrian movement but also accurately predict the pedestrian movement trajectory.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Adapt2Reward: Adapting Video-Language Models to Generalizable Robotic Rewards via Failure Prompts

arXiv

引用

arXiv 2024年

作者： Yang, Yanting Chen, Minghao Qiu, Qibo Wu, Jiahao Wang, Wenxiao Lin, Binbin Guan, Ziyu He, Xiaofei School of Software Technology Zhejiang University China School of Computer Sciene and Technology Hangzhou Dianzi University China Research & Innovation Institute China The Hong Kong Polytechnic University Hong Kong Zhiyuan Research Institute China School of Computer Sciene and Technology Xidian University China State Key Lab of CAD&CG Zhejiang University China

For a general-purpose robot to operate in reality, executing a broad range of instructions across various environments is imperative. Central to the reinforcement learning and planning for such robotic agents is a generalizable reward function. Recent advances in vision-language models, such as CLIP, have shown remarkable performance in the domain of deep learning, paving the way for open-domain visual recognition. However, collecting data on robots executing various language instructions across multiple environments remains a challenge. This paper aims to transfer video-language models with robust generalization into a generalizable language-conditioned reward function, only utilizing robot video data from a minimal amount of tasks in a singular environment. Unlike common robotic datasets used for training reward functions, human video-language datasets rarely contain trivial failure videos. To enhance the model’s ability to distinguish between successful and failed robot executions, we cluster failure video features to enable the model to identify patterns within. For each cluster, we integrate a newly trained failure prompt into the text encoder to represent the corresponding failure mode. Our language-conditioned reward function shows outstanding generalization to new environments and new instructions for robot planning and reinforcement learning. © 2024, CC BY.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

CEGRL-TKGR: A Causal Enhanced Graph Representation Learning Framework for Temporal Knowledge Graph Reasoning

arXiv

引用

arXiv 2024年

作者： Sun, Jinze Sheng, Yongpan He, Lirong Qin, Yongbin Liu, Ming Jia, Tao College of Computer and Information Science Southwest University Chongqing400715 China School of Big Data & Software Engineering Chongqing University Chongqing401331 China School of Information Science and Engineering Chongqing Jiaotong University Chongqing400074 China College of Computer Science and Technology The State Key Laboratory of Public Big Data Guizhou University China Deakin University VIC3125 Australia

Temporal knowledge graph reasoning (TKGR) is increasingly gaining attention for its ability to extrapolate new events from historical data, thereby enriching the inherently incomplete temporal knowledge graphs. Existing graph-based representation learning frameworks have made significant strides in developing evolving representations for both entities and relational embeddings. Despite these achievements, there’s a notable tendency in these models to inadvertently learn biased data representations and mine spurious correlations, consequently failing to discern the causal relationships between events. This often leads to incorrect predictions based on these false correlations. To address this, we propose an innovative Causal Enhanced Graph Representation Learning framework for TKGR (named CEGRL-TKGR). This framework introduces causal structures in graph-based representation learning to unveil the essential causal relationships between events, ultimately enhancing the performance of the TKGR task. Specifically, we first disentangle the evolutionary representations of entities and relations in a temporal knowledge graph sequence into two distinct components, namely causal representations and confounding representations. Then, drawing on causal intervention theory, we advocate the utilization of causal representations for predictions, aiming to mitigate the effects of erroneous correlations caused by confounding features, thus achieving more robust and accurate predictions. Finally, extensive experimental results on six benchmark datasets demonstrate the superior performance of our model in the link prediction task. © 2024, CC BY.

关键词： Spatio-temporal data

来源：评论

学校读者我要写书评

暂无评论

Robust Zero Level-Set Extraction from Unsigned Distance Fields Based on Double Covering

arXiv

引用

arXiv 2023年

作者： Hou, Fei Chen, Xuhui Wang, Wencheng Qin, Hong He, Ying State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences University of Chinese Academy of Sciences Beijing China Department of Computer Science Stony Brook University New York United States School of Computer Science and Engineering Nanyang Technological University Singapore

In this paper, we propose a new method, called DoubleCoverUDF, for extracting the zero level-set from unsigned distance fields (UDFs). DoubleCoverUDF takes a learned UDF and a user-specified parameter r (a small positive real number) as input and extracts an iso-surface with an iso-value r using the conventional marching cubes algorithm. We show that the computed iso-surface is the boundary of the r-offset volume of the target zero level-set S, which is an orientable manifold, regardless of the topology of S. Next, the algorithm computes a covering map to project the boundary mesh onto S, preserving the mesh's topology and avoiding folding. If S is an orientable manifold surface, our algorithm separates the double-layered mesh into a single layer using a robust minimum-cut post-processing step. Otherwise, it keeps the double-layered mesh as the output. We validate our algorithm by reconstructing 3D surfaces of open models and demonstrate its efficacy and effectiveness on synthetic models and benchmark datasets. Our experimental results confirm that our method is robust and produces meshes with better quality in terms of both visual evaluation and quantitative measures than existing UDF-based methods. The source code is available at https://***/jjjkkyz/DCUDF. © 2023, CC BY-NC-ND.

关键词： Topology

来源：评论

学校读者我要写书评

暂无评论

Corrosion Detection in Transformers Based on Hierarchical Annotation 7th

Corrosion Detection in Transformers Based on Hierarchical An...

引用

7th International Conference on Artificial Intelligence and Security, ICAIS 2021

作者： Cao, Yong Zhou, Yinian Zhang, Zhao Wu, Wenjun Chen, Xihai Yang, Sha Zhang, Baili School of Computer Science and Engineering Southeast University Nanjing China School of Computer and Software Nanjing University of Information Science and Technology Nanjing China State Key Laboratory of Smart Grid Protection and Control Nanjing China Nari Group Corporation Nanjing China

ISBN: (纸本)9783030786083

Transformers are prone to metal corrosion due to long-term exposure in the external environment. It is necessary to detect corrosion in time to prolong their service life-time. Detecting metal corrosion based on deep learning is becoming a feasible method at present. However, because of the irregularity and detachability of metal corrosion, traditional annotation approach results in such problems as ambiguity and uncertainty in the annotating process. Thus, a novel hierarchical annotation approach is proposed in this paper. The rationale for this initiative is as follows: traditional annotation approach has been used to annotate a large area covering the range of corrosion, as long as the area is visually continuous and adjacent to the corrosion that cannot be clearly divided. The annotation result is recorded as training set A;secondly, in the annotating boxes from the first step, the areas with obvious and relatively independent features are re-annotated to form the second level of nested annotation. The annotation result is recorded as training set B. Finally, Faster R-CNN and YOLOv5 models were trained by A, B training sets respectively in the experiment. It can be concluded that the detection performance with hierarchical annotation approach proved better than traditional annotation approach for Faster R-CNN and YOLOv5 models. © 2021, Springer Nature Switzerland AG.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Hyperspectral Image-Text Coupling Network For Hyperspectral Image Classification

Hyperspectral Image-Text Coupling Network For Hyperspectral ...

引用

IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)

作者： Jiaqi Yang Bo Du State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing Wuhan University Wuhan P. R. China National Engineering Research Center for Multimedia Software Institute of Artificial Intelligence School of Computer Science and Hubei Key Laboratory of Multimedia and Network Communication Engineering Wuhan University Wuhan P. R. China

ISBN: (数字)9798350360325

ISBN: (纸本)9798350360332

Deep learning-based approaches have blossomed in the field of hyperspectral image (HSI) classification. However, most of the existing methods focus only on visual information and ignore textual clues. Textual properties can describe the shape, color, and size of ground objects with a wealth of information and are a great complement to visual features. Therefore, the HSI-Text (HSI-T) coupling network is proposed in this study. Specifically, a geocoder is first designed to encode the geospatial information of the image. Moreover, a textual transformer is introduced to extract the attributes of the language. On the basis of the above structures, an HSI-T coupling network is built to explicitly model visual-textual dependencies. Experimental results on benchmark HSI datasets can verify that the proposed approach can gain better classification accuracy and more fine-grained land cover maps.

关键词： Couplings Visualization Shape Image color analysis Land surface Transformers Feature extraction

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：