检索结果-内蒙古大学图书馆

Empirical studies on the properties of linear regions in deep neural networks

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Zhang, Xiao Wu, Dongrui Ministry of Education Key Laboratory of Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China

A deep neural network (DNN) with piecewise linear activations can partition the input space into numerous small linear regions, where different linear functions are fitted. It is believed that the number of these regions represents the expressivity of the DNN. This paper provides a novel and meticulous perspective to look into DNNs: Instead of just counting the number of the linear regions, we study their local properties, such as the inspheres, the directions of the corresponding hyperplanes, the decision boundaries, and the relevance of the surrounding regions. We empirically observed that different optimization techniques lead to completely different linear regions, even though they result in similar classification accuracies. We hope our study can inspire the design of novel optimization techniques, and help discover and analyze the behaviors of DNNs. Copyright © 2020, The Authors. All rights reserved.

关键词： Deep neural networks

Integrating Informativeness, Representativeness and Diversity in Pool-Based Sequential Active Learning for Regression

学校读者我要写书评

暂无评论

Integrating Informativeness, Representativeness and Diversit...

International Joint Conference on Neural Networks (IJCNN)

作者： Ziang Liu Dongrui Wu Ministry of Education Key Laboratory of Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China

ISBN: (数字)9781728169262

ISBN: (纸本)9781728169279

In many real-world machine learning applications, unlabeled samples are easy to obtain, but it is expensive and/or time-consuming to label them. Active learning is a common approach for reducing this data labeling effort. It optimally selects the best few samples to label, so that a better machine learning model can be trained from the same number of labeled samples. This paper considers active learning for regression (ALR) problems. Three essential criteria - informativeness, representativeness, and diversity - have been proposed for ALR. However, very few approaches in the literature have considered all three of them simultaneously. We propose three new ALR approaches, with different strategies for integrating the three criteria. Extensive experiments on 12 datasets in various domains demonstrated their effectiveness.

关键词： Computational modeling Clustering algorithms Machine learning Training Labeling Predictive models image processing

Down-Level control Based on Level Prediction of Landslide Evolutionary State

学校读者我要写书评

暂无评论

Down-Level Control Based on Level Prediction of Landslide Ev...

International Conference on Information Science and Technology (ICIST)

作者： Shu Sun Cheng Lian Xiaoping Wang School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China School of Automation Wuhan University of Technology Wuhan China School of Artificial Intelligence and Automation Key Laboratory of Image Processing and Intelligent Control Huazhong University of Science and Technology Wuhan China

Accurate prediction and effective control are the key to reducing the impact of landslide disasters. In practice, most of recent studies focus on landslide displacement prediction and control. In this paper, we propose a new control method based on the level prediction of landslide evolution state, namely down-level control. The core components are level predictor and interval predictor in this method. Specially, we first transform the traditional displacement regression prediction problem into a level classification prediction problem by labeling discrete category information for landslide sample points. Then the level predictor based on Multi-task learning-Stacked long-short time memory network (MTL-SLSTM) is established to predict the state of the landslide and judge whether the system needs to be controlled. Finally, we design a safe rainfall interval predictor based on bootstrap method to obtain the safe value of control variation. The effectiveness of the proposed control method is verified on Baishuihe and Shiliushubao landslides. The results show the proposed down-level control method is valid and more intuitive.

关键词： Information science Geology Transforms Predictive models Terrain factors Complexity theory Labeling

NiuEM: A Nested-iterative Unsupervised Learning Model for Single-particle Cryo-EM image processing

学校读者我要写书评

暂无评论

NiuEM: A Nested-iterative Unsupervised Learning Model for Si...

2020 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2020

作者： Hu, Rui Cai, Jiaming Zheng, Wangjie Yang, Yang Shen, Hong-Bin Shanghai Jiao Tong University Key Lab. of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Department of Computer Science and Engineering Shanghai200240 China Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai200240 China Shanghai Jiao Tong University Department of Bioinformatics and Biostatistics Shanghai200240 China

ISBN: (纸本)9781728162157

Cryo-electron microscopy (cryo-EM) has become a mainstream technology for solving spatial structures of biomacromolecules, while the processing of cryo-EM images is a very challenging task. One of the great challenges is the high noise in the images. A common method is to cluster the images with close projecting angles to get mean images, which are used for 3D reconstruction. However, due to the extremely low signal-to-noise-ratio, common clustering methods often fail to obtain high-quality mean images, leading to poorly reconstructed structures. In this study, we present a new unsupervised learning framework, called NiuEM, to discriminate images captured from different angles and yield cluster-mean images. NiuEM first generates pseudo-labels and then exploits both contrastive loss and cross-entropy loss for training convolutional layers to learn feature representations. Moreover, the pseudo-labels are updated iteratively to enhance the reliability of labels. We assess the performance of NiuEM on four data sets via both visualized and quantitative experiments. Especially, two kinds of metrics are adopted to measure the performance, regarding the clustering quality and the resolution of reconstructed 3D models, respectively. The experimental results show that NiuEM achieves very competitive clustering accuracy in the comparison with the state-of-the-art image clustering methods. Moreover, the cluster mean images yielded by NiuEM lead to better initial 3D models compared with the mainstream reconstruction tools. © 2020 IEEE.

关键词： image reconstruction

One-shot distributed algorithm for PCA with RBF kernels

学校读者我要写书评

暂无评论

arXiv 2020年

作者： He, Fan Lv, Kexin Yang, Jie Huang, Xiaolin The Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University China The MOE Key Laboratory of System Control and Information Processing 800 Dongchuan Road Shanghai200240 China

This letter proposes a one-shot algorithm for feature-distributed kernel PCA. Our algorithm is inspired by the dual relationship between sample-distributed and feature-distributed scenario. This interesting relationship makes it possible to establish distributed kernel PCA for feature-distributed cases from ideas in distributed PCA in sample-distributed scenario. In theoretical part, we analyze the approximation error for both linear and RBF kernels. The result suggests that when eigenvalues decay fast, the proposed algorithm gives high quality results with low communication cost. This result is also verified by numerical experiments, showing the effectiveness of our algorithm in practice. Copyright © 2020, The Authors. All rights reserved.

关键词： Principal component analysis

Combining Attractor Propagation with Differential Evolution for Three-echelon logistics distribution optimization

学校读者我要写书评

暂无评论

SSRN

SSRN 2022年

作者： Zhang, Haifei Ge, Hongwei Yang, Jieming Su, Shuzhi Tong, Yubing School of Artificial Intelligence and Computer Science Jiangnan University Jiangsu Wuxi214122 China Key Laboratory of Advanced Process Control for Light Industry Jiangnan University Ministry of Education Wuxi214122 China School of Computer Science and Engineering Anhui University of Science & Technology Anhui Huainan232001 China Medical Image Processing Group Department of Radiology University of Pennsylvania PhiladelphiaPA19104 United States

In the process of urban logistics distribution, how to achieve efficiently and timely logistics distribution within the distribution time limit has become an urgent problem to be solved in logistics management. Three-echelon logistics distribution is considered to be an effective strategy to alleviate urban congestion, improve liquidity and reduce pollution. In this paper, a mathematical model of three-echelon logistics distribution problem considering time window constraint is established based on directed graph, and a double-tier intelligent algorithm coordinated by Distance Entropy-based Attractor Propagation (DEBAP) clustering algorithm - Crossover-and-Selection based Differential Evolution (CSBDE) is proposed. DEBAP is responsible for the optimization of region partitioning and facility allocation. The obtained region partitioning and facility allocation scheme are transmitted to CSBDE for path planning. CSBDE feeds back the optimized path planning scheme to DEBAP, and the two-tier intelligent algorithm cooperates to complete the optimization of three-echelon logistics distribution system. In addition, this paper proposes a search strategy based on reachable distribution region and a facility allocation optimization strategy based on path length as weight to improve the efficiency of the algorithm. The example calculation shows that the mathematical model and combined intelligent algorithm proposed in this paper can effectively solve 3-echelon logistics distribution problem(3E-LDP), provide theoretical and methodological guidance for multi-echelon logistics distribution, and can be well implemented in practice to help logistics operators reduce operation costs and improve customer service. © 2022, The Authors. All rights reserved.

关键词： Clustering algorithms

MBGD-RDA Training and Rule Pruning for Concise TSK Fuzzy Regression Models

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wu, Dongrui Key Laboratory of the Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

To effectively train Takagi-Sugeno-Kang (TSK) fuzzy systems for regression problems, a Mini-Batch Gradient Descent with Regularization, DropRule, and AdaBound (MBGD-RDA) algorithm was recently proposed. It has demonstrated superior performances;however, there are also some limitations, e.g., it does not allow the user to specify the number of rules directly, and only Gaussian MFs can be used. This paper proposes two variants of MBGD-RDA to remedy these limitations, and show that they outperform the original MBGD-RDA and the classical ANFIS algorithms with the same number of rules. Furthermore, we also propose a rule pruning algorithm for TSK fuzzy systems, which can reduce the number of rules without significantly sacrificing the regression performance. Experiments showed that the rules obtained from pruning are generally better than training them from scratch directly, especially when Gaussian MFs are used. Copyright © 2020, The Authors. All rights reserved.

关键词： Gradient methods

Multi-Level temporal pyramid network for action detection

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wang, Xiang Gao, Changxin Zhang, Shiwei Sang, Nong Key Laboratory of Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology China DAMO Academy Alibaba Group

Currently, one-stage frameworks have been widely applied for temporal action detection, but they still suffer from the challenge that the action instances span a wide range of time. The reason is that these one-stage detectors, e.g., Single Shot Multi-Box Detector (SSD), extract temporal features only applying a single-level layer for each head, which is not discriminative enough to perform classification and regression. In this paper, we propose a Multi-Level Temporal Pyramid Network (MLTPN) to improve the discrimination of the features. Specially, we first fuse the features from multiple layers with different temporal resolutions, to encode multi-layer temporal information. We then apply a multi-level feature pyramid architecture on the features to enhance their discriminative abilities. Finally, we design a simple yet effective feature fusion module to fuse the multi-level multi-scale features. By this means, the proposed MLTPN can learn rich and discriminative features for different action instances with different durations. We evaluate MLTPN on two challenging datasets: THUMOS14 and Activitynet v1.3, and the experimental results show that MLTPN obtains competitive performance on Activitynet v1.3 and outperforms the state-of-the-art approaches on THUMOS14 significantly. Copyright © 2020, The Authors. All rights reserved.

关键词： Feature extraction

Single upper limb pose estimation method based on improved stacked hourglass network

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Peng, Gang Zheng, Yuezhi Li, Jianfeng Yang, Jin Deng, Zhonghua Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

At present, most high-accuracy single-person pose estimation methods have high computational complexity and insufficient real-time performance due to the complex structure of the network model. However, a single-person pose estimation method with high real-time performance also needs to improve its accuracy due to the simple structure of the network model. It is currently difficult to achieve both high accuracy and real-time performance in single-person pose estimation. For use in human–machine cooperative operations, this paper proposes a single-person upper limb pose estimation method based on an end-to-end approach for accurate and real-time limb pose estimation. Using the stacked hourglass network model, a single-person upper limb skeleton key point detection model was designed. Deconvolution was employed to replace the up-sampling operation of the hourglass module in the original model, solving the problem of rough feature maps. Integral regression was used to calculate the position coordinates of key points of the skeleton, reducing quantization errors and calculations. Experiments showed that the developed single-person upper limb skeleton key point detection model achieves high accuracy and that the pose estimation method based on the end-to-end approach provides high accuracy and real-time performance. Copyright © 2020, The Authors. All rights reserved.

关键词： Convolutional neural networks