检索结果-内蒙古大学图书馆

arXiv 2021年

作者： Li, Qiufu Shen, Linlin Guo, Sheng Lai, Zhihui Computer Vision Institute College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen518060 China Guangdong Key Laboratory of Intelligent Information Processing Shenzhen University Shenzhen518060 China MyBank Ant Group Hangzhou310012 China

Though widely used in image classification, convolutional neural networks (CNNs) are prone to noise interruptions, i.e. the CNN output can be drastically changed by small image noise. To improve the noise robustness, we try to integrate CNNs with wavelet by replacing the common down-sampling (maxpooling, strided-convolution, and average pooling) with discrete wavelet transform (DWT). We firstly propose general DWT and inverse DWT (IDWT) layers applicable to various orthogonal and biorthogonal discrete wavelets like Haar, Daubechies, and Cohen, etc., and then design wavelet integrated CNNs (WaveCNets) by integrating DWT into the commonly used CNNs (VGG, ResNets, and DenseNet). During the down-sampling, WaveCNets apply DWT to decompose the feature maps into the low-frequency and high-frequency components. Containing the main information including the basic object structures, the low-frequency component is transmitted into the following layers to generate robust high-level features. The high-frequency components are dropped to remove most of the data noises. The experimental results show that WaveCNets achieve higher accuracy on ImageNet than various vanilla CNNs. We have also tested the performance of WaveCNets on the noisy version of ImageNet, ImageNet-C and six adversarial attacks, the results suggest that the proposed DWT/IDWT layers could provide better noise-robustness and adversarial robustness. When applying WaveCNets as backbones, the performance of object detectors (i.e., faster R-CNN and RetinaNet) on COCO detection dataset are consistently improved. We believe that suppression of aliasing effect, i.e. separation of low frequency and high frequency information, is the main advantages of our approach. The code of our DWT/IDWT layer and differentWaveCNets are available at https://***/CVISZU/ WaveCNet. © 2021, CC BY.

关键词： Signal sampling

来源：评论

学校读者我要写书评

暂无评论

Efficient TD3 based path planning of mobile robot in dynamic environments using prioritized experience replay and LSTM

引用

Scientific reports 2025年第1期15卷 18331页

作者： Yunhan Lin Zhijie Zhang Yijian Tan Hao Fu Huasong Min School of Computer Science and Technology Wuhan University of Science and Technology Wuhan 430081 China. Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System Wuhan University of Science and Technology Wuhan 430081 China. Institute of Robotics and Intelligent Systems Wuhan University of Science and Technology Wuhan 430081 China. Institute of Robotics and Intelligent Systems Wuhan University of Science and Technology Wuhan 430081 China. mhuasong@***.

To address the challenges of sample utilization efficiency and managing temporal dependencies, this paper proposes an efficient path planning method for mobile robot in dynamic environments based on an improved twin delayed deep deterministic policy gradient (TD3) algorithm. The proposed method, named PL-TD3, integrates prioritized experience replay (PER) and long short-term memory (LSTM) neural networks, which enhance both sample efficiency and the ability to handle time-series data. To verify the effectiveness of the proposed method, simulation and practical experiments were designed and conducted. In the simulation experiments, both static and dynamic obstacles were included in the test environment, along with experiments to assess generalization capabilities. The algorithm demonstrated superior performance in terms of both execution time and path efficiency. The practical experiments, based on the assumptions from the simulation tests, further confirmed that PL-TD3 has improved the effectiveness and robustness of path planning for mobile robot in dynamic environments.

关键词： Long short-term memory (LSTM ) Path planning in dynamic environment Prioritized experience replay (PER) Reinforcement learning Twin delayed deep deterministic policy gradient (TD3) algorithm

来源：评论

学校读者我要写书评

暂无评论

An Order Dispatch System Based on Reinforcement Learning for Ride Sharing Services

An Order Dispatch System Based on Reinforcement Learning for...

引用

IEEE International Conference on High Performance Computing and Communications (HPCC)

作者： Zeqiang Chen Peng Li Junlei Xiao Lei Nie Yu Liu College of Computer Science and Technology Wuhan University of Science and Technology China Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System China

ISBN: (纸本)9781728176499;9781728176505

Ride-sharing has been widely used in many cities, such as Didi and Uber. Ride-sharing is regarded as an effective way to solve urban traffic congestion and pollution. However, most of the existing dispatch methods take the minimization of the travel distance as the optimization goal, without considering other factors. In this paper, we consider not only detour distance, but also consider seat utilization, future profit, and hidden profit. Firstly, we propose a deep evaluation network to evaluate factors that affect vehicle dispatch, and we exploit the reinforcement learning strategy to train the deep evaluation network. Then, we propose the dynamic external factor calculation (DEFC) algorithm to calculate those factors. Secondly, to calculate the driver's future profit, we propose a prediction model based on the K-Nearest Neighbor(KNN) to predict the number of passengers and vehicles. Thirdly, we use a vehicle search method to search vehicles that satisfy passenger spatial-temporal constraint. Based on a real-world dataset in Rome, we evaluate our algorithm to confirm the effectiveness of our method.

关键词： Search methods Heuristic algorithms Urban areas Reinforcement learning Predictive models Prediction algorithms Dispatching

来源：评论

学校读者我要写书评

暂无评论

Data Analysis and Planning Analysis Management System Based on Tensorflow Machine Learning Algorithms

Data Analysis and Planning Analysis Management System Based ...

引用

2019 International Conference on Artificial Intelligence Technologies and Applications, ICAITA 2019

作者： Hu, Shuang Huang, Li Bai, Feng Zhang, Jun School of Computer Science and Technology Wuhan University of Science and Technology China Intelligent Information Processing and Real-time Industrial System Hubei Key Laboratory Institute China

At present, machine learning is quite popular, and tensorflow framework is also very popular. This system takes the plan management system as the practice platform, realizes the data analysis function through the machine learning understanding and the key point. Most of the current market are planning management system, but there is no self-analysis ability, need to base on data, there are limitations. Firstly, the system collects training data such as weight, body fat rate and so on, then builds a deep learning neural network, and finally runs the model to realize plan analysis, so as to help users analyze the feasibility of the plan and provide suggestions for users, and the analysis results will be displayed through processing. © 2019 IOP Publishing Ltd. All rights reserved.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Research on Question Answering System Based on Bi-LSTM and Self-attention Mechanism

Research on Question Answering System Based on Bi-LSTM and S...

引用

International Conference on Industrial Engineering and Applications (ICIEA)

作者： Hao Xiang Jinguang Gu School of Computer Science and Technology Wuhan University of Science and Technology Key Laboratory of Intelligent Information Processing and Real-time Industrial System in Hubei Province Wuhan China

ISBN: (数字)9781728167855

ISBN: (纸本)9781728167862

With the development of artificial intelligence technology, intelligent question answering has become a hot research direction in the field of natural language processing. This paper proposes a question answering method based on Bi-LSTM and self-attention mechanism model. This method uses Bi-LSTM to encode and align the question and answer respectively, then uses self-attention to obtain the relationship between keywords, and finally performs softmax through the fully connected layer to obtain the similarity between the question and answer. Finally, in the experiment, compared with the traditional attention model, the accuracy rate of this model was increased by 1.6%, and the recall rate was increased by 1.5%.

关键词： Semantics Machine learning Task analysis Knowledge discovery Feature extraction Standards Training

来源：评论

学校读者我要写书评

暂无评论

Deformation Flow Based Two-Stream Network for Lip Reading

Deformation Flow Based Two-Stream Network for Lip Reading

引用

International Conference on Automatic Face and Gesture recognition

作者： Jingyun Xiao Shuang Yang Yuanhang Zhang Shiguang Shan Xilin Chen Key Laboratory of Intelligent Information Processing of Chinese Academy of Sciences (CAS) Institute of Computing Technology CAS Beijing China University of Chinese Academy of Sciences Beijing China

ISBN: (数字)9781728130798

ISBN: (纸本)9781728130804

Lip reading is the task of recognizing speech content by analyzing movements in the lip region when people are speaking. Based on the continuity in adjacent frames in the speaking process, and the consistency in motion patterns among different people when they pronounce the same phoneme, we model lip movements as a sequence of apparent deformations in the lip region during the speaking process. Specifically, we introduce a Deformation Flow Network (DFN) to learn the deformation flow between adjacent frames, which directly captures the motion information within the lip region. The learned deformation flow is then combined with the original grayscale frames with a two-stream network to perform lip reading. To make the two streams learn from each other in the learning process, we introduce a bidirectional knowledge distillation loss to train the two branches jointly. Owing to the complementary cues provided by different branches, the two-stream network shows substantial improvement over using either single branch. A thorough experimental evaluation on two large-scale lip reading benchmarks is presented with detailed analysis. The results accord with our motivation, and show that our method achieves state-of-the-art or comparable performance on these two challenging datasets.

关键词： Strain Lips Face recognition Task analysis Training Decoding Gray-scale

来源：评论

学校读者我要写书评

暂无评论

M3F: Multi-Modal Continuous Valence-Arousal Estimation in the Wild

M3F: Multi-Modal Continuous Valence-Arousal Estimation in th...

引用

International Conference on Automatic Face and Gesture recognition

作者： Yuan-Hang Zhang Rulin Huang Jiabei Zeng Shiguang Shan Key Laboratory of Intelligent Information Processing of Chinese Academy of Sciences (CAS) Institute of Computing Technology CAS Beijing China University of Chinese Academy of Sciences Beijing China

ISBN: (数字)9781728130798

ISBN: (纸本)9781728130804

In this paper, we propose a multi-modal multi-feature (M$^{3}F$) approach for in-the-wild valence-arousal estimation. In the proposed M$^{3}F$ framework, we fuse both visual features from videos and acoustic features from the audio tracks to estimate the valence and arousal. We follow a CNN-RNN paradigm, where the spatio-temporal visual features are extracted with a 3D convolutional network and/or a pretrained 2D convolutional network, and a bidirectional recurrent neural network. We evaluated the M$^{3}F$ framework on the validation set provided by the Affective Behavior Analysis in-the-wild (ABAW) Challenge, held in conjunction with the IEEE International Conference on Automatic Face and Gesture recognition (FG) 2020, and it significantly outperforms the baseline method.

关键词： Feature extraction Visualization Three-dimensional displays Two dimensional displays Videos Training Estimation

来源：评论

学校读者我要写书评

暂无评论

A comprehensive assessment method of asset importance for key asset protection 9

A comprehensive assessment method of asset importance for ke...

引用

9th International Conference on Communication and Network Security, ICCNS 2019

作者： Gao, Xiaoling Liu, Yanhua Qiu, Yanbin Liu, Huajun College of Mathematics and Computer Science Fuzhou University Fuzhou Fujian China Fujian Provincial Key Laboratory of Network Computing and Intelligent Information Processing Fuzhou Fujian China

ISBN: (纸本)9781450376624

In a given system network, an important prerequisite for security risk control is how to accurately calculate the impact of different host assets in the topology on the spread of attack risk. In this regard, we propose a comprehensive asset importance assessment model for critical asset protection. First, the importance of local static assets in the topology was evaluated using three methods: degree centrality, mediation centrality, and closeness centrality. At the same time, PageRank was used to assess the importance of local dynamic assets for asset level changes. By combining the above two methods, and using K-CORE to evaluate the core position of the host in the network topology as the global weight indicator of the importance of assets. The comprehensive evaluation results of the importance of topology-oriented assets was finally obtained, and the validity of the model was verified through the experiment of the simulation ER random network. © 2019 Association for Computing Machinery.

关键词： Network topology

来源：评论

学校读者我要写书评

暂无评论

Ultrasound vascular image segmentation based on full convolutional aggregation neural network 3

Ultrasound vascular image segmentation based on full convolu...

引用

3rd International Conference on Computer Science and Artificial Intelligence, CSAI 2019 and its Workshop - The 11th International Conference on information and Multimedia Technology, ICIMT 2019

作者： Li, Yan Liu, Jun Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial Department of Computer Science and Technology Wuhan University of Science and technology Wuhan430070 China

ISBN: (纸本)9781450376273

The concept of ultrasound therapy has been proposed long ago. However, most of the previous methods of ultrasound therapy destroy tissue through thermal effects and cause great damage to patients. In recent years, ultrasound cavitation therapy has caused extensive discussion and research due to its unique noninvasiveness. In order to achieve real-time access to the patient during treatment, the lesion area must be processed simultaneously and accurately. However, ultrasound images mostly have lower signal-to-noise ratio, contrast, and blurred edges. In order to solve the segmentation problem, this paper proposes a new segmentation model Du-net based on the full convolutional neural network. Under the premise of deepening the network depth to obtain more information, the Encoder- Decoder method and the layered aggregation mode are used to prevent the gradient explosion. Solving boundary segmentation problem by constructing joint network, In the case of a small data set, the data enhancement method using a suitable data set effectively increases the usable image features and achieves better results on the ultrasound blood vessel image. © 2019 Association for Computing Machinery.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Inverse linear-quadratic discrete-time finite-horizon optimal control for indistinguishable homogeneous agents: a convex optimization approach

arXiv

引用

arXiv 2021年

作者： Zhang, Han Ringh, Axel Department of Automation School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai200240 China Shanghai Engineering Research Center of Intelligent Control and Management Shanghai200240 China Department of Mathematical Sciences Chalmers University of Technology University of Gothenburg Gothenburg41296 Sweden

The inverse linear-quadratic optimal control problem is a system identification problem whose aim is to recover the quadratic cost function and hence the closed-loop system matrices based on observations of optimal trajectories. In this paper, the discrete-time, finite-horizon case is considered, where the agents are also assumed to be homogeneous and indistinguishable. The latter means that the agents all have the same dynamics and objective functions and the observations are in terms of "snap shots" of all agents at different time instants, but what is not known is "which agent moved where" for consecutive observations. This absence of linked optimal trajectories makes the problem challenging. We first show that this problem is globally identifiable. Then, for the case of noiseless observations, we show that the true cost matrix, and hence the closed-loop system matrices, can be recovered as the unique global optimal solution to a convex optimization problem. Next, for the case of noisy observations, we formulate an estimator as the unique global optimal solution to a modified convex optimization problem. Moreover, the statistical consistency of this estimator is shown. Finally, the performance of the proposed method is demonstrated by a number of numerical examples. Copyright © 2021, The Authors. All rights reserved.

关键词： Convex optimization

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：