检索结果-内蒙古大学图书馆

MetAug: Contrastive Learning via Meta Feature Augmentation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Li, Jiangmeng Qiang, Wenwen Zheng, Changwen Su, Bing Xiong, Hui Science & Technology on Integrated Information System Laboratory Institute of Software Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Guangdong China Gaoling School of Artificial Intelligence Renmin University of China Beijing China Beijing Key Laboratory of Big Data Management and Analysis Methods Beijing China Guangzhou China Department of Computer Science Engineering The Hong Kong University of Science and Technology Hong Kong

What matters for contrastive learning? We argue that contrastive learning heavily relies on informative features, or "hard" (positive or negative) features. Early works include more informative features by applying complex data augmentations and large batch size or memory bank, and recent works design elaborate sampling approaches to explore informative features. The key challenge toward exploring such features is that the source multi-view data is generated by applying random data augmentations, making it infeasible to always add useful information in the augmented data. Consequently, the informativeness of features learned from such augmented data is limited. In response, we propose to directly augment the features in latent space, thereby learning discriminative representations without a large amount of input data. We perform a meta learning technique to build the augmentation generator that updates its network parameters by considering the performance of the encoder. However, insufficient input data may lead the encoder to learn collapsed features and therefore malfunction the augmentation generator. A new margin-injected regularization is further added in the objective function to avoid the encoder learning a degenerate mapping. To contrast all features in one gradient back-propagation step, we adopt the proposed optimization-driven unified contrastive loss instead of the conventional contrastive loss. Empirically, our method achieves state-of-the-art results on several benchmark datasets. Copyright © 2022, The Authors. All rights reserved.

关键词： Backpropagation

Multimodal Collaboration Networks for Geospatial Vehicle Detection in Dense, Occluded, and Large-Scale Events

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wu, Xin Huang, Zhanchao Wang, Li Chanussot, Jocelyn Tian, Jiaojiao The School of Computer Science National Pilot Software Engineering School Beijing University of Posts and Telecommunications Beijing100876 China The Key Laboratory of Spatial Data Mining and Information Sharing The Ministry of Education The Academy of Digital China Fuzhou University Fuzhou350108 China The National & Local Joint Engineering Research Center of Satellite Geospatial Information Technology Fuzhou University Fuzhou350108 China The Univ. Grenoble Alpes CNRS Grenoble INP GIPSA-Lab Grenoble38000 France The Aerospace Information Research Institute Chinese Academy of Sciences Beijing100094 China The Remote Sensing Technology Institute German Aerospace Center Weßling82205 Germany

In large-scale disaster events, the planning of optimal rescue routes depends on the object detection ability at the disaster scene, with one of the main challenges being the presence of dense and occluded objects. Existing methods, which are typically based on the RGB modality, struggle to distinguish targets with similar colors and textures in crowded environments and are unable to identify obscured objects. To this end, we first construct two multimodal dense and occlusion vehicle detection datasets for large-scale events, utilizing RGB and height map modalities. Based on these datasets, we propose a multimodal collaboration network for dense and occluded vehicle detection, MuDet for short. MuDet hierarchically enhances the completeness of discriminable information within and across modalities and differentiates between simple and complex samples. MuDet includes three main modules: Unimodal Feature Hierarchical Enhancement (Uni-Enh), Multimodal Cross Learning (Mul-Lea), and Hard-easy Discriminative (He-Dis) Pattern. Uni-Enh and Mul-Lea enhance the features within each modality and facilitate the cross-integration of features from two heterogeneous modalities. He-Dis effectively separates densely occluded vehicle targets with significant intra-class differences and minimal inter-class differences by defining and thresholding confidence values, thereby suppressing the complex background. Experimental results on two re-labeled multimodal benchmark datasets, the 4K-SAI-LCS dataset, and the ISPRS Potsdam dataset, demonstrate the robustness and generalization of the MuDet. The codes of this work are available openly at https://***/Shank2358/MuDet. © 2024, CC BY-NC-SA.

关键词： Remote sensing

Attention mechanism meets with hybrid dense network for Hyperspectral Image Classification

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Ahmad, Muhammad Mehmood Khan, Adil Mazzara, Manuel Distefano, Salvatore Roy, Swalpa Kumar Wu, Xin Department of Computer Science National University of Computer and Emerging Sciences Islamabad Chiniot-Faisalabad Campus Chiniot35400 Pakistan Institute of Data Science and Artificial Intelligence Innopolis University Innopolis420500 Russia Institute of Software Development and Engineering Innopolis University Innopolis420500 Russia Dipartimento di Matematica e Informatica MIFT University of Messina Messina98121 Italy Department of Computer Science and Engineering Jalpaiguri Government Engineering College West Bengal 735102 India School of Information and Electronics Beijing Institute of Technology Beijing100081 China

The nonlinear relation between the spectral information and the corresponding objects (complex physiognomies) makes pixel-wise classification challenging for conventional methods. To deal with nonlinearity issues in Hyperspectral Image Classification (HISC), Convolutional Neural Networks (CNN) are more suitable, indeed. However, fixed kernel sizes make traditional CNN too specific, neither flexible nor conducive to feature learning, thus impacting on the classification accuracy. The convolution of different kernel size networks may overcome this problem by capturing more discriminating and relevant information. In light of this, the proposed solution aims at combining the core idea of 3D and 2D Inception net with the Attention mechanism to boost the HSIC CNN performance in a hybrid scenario. The resulting attention-fused hybrid network (AfNet) is based on three attention-fused parallel hybrid subnets with different kernels in each block repeatedly using highlevel features to enhance the final ground-truth maps. In short, AfNet is able to selectively filter out the discriminative features critical for classification. Several tests on HSI datasets provided competitive results for AfNet compared to state-of-the-art models. The proposed pipeline achieved, indeed, an overall accuracy of 97% for the Indian Pines, 100% for Botswana, 99% for Pavia University, Pavia Center, and Salinas datasets. © 2022, CC BY.

关键词： Convolution

Interventional Contrastive Learning with Meta Semantic Regularizer

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Qiang, Wenwen Li, Jiangmeng Zheng, Changwen Su, Bing Xiong, Hui Science & Technology on Integrated Information System Laboratory Institute of Software Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Guangdong China Gaoling School of Artificial Intelligence Renmin University of China Beijing China Beijing Key Laboratory of Big Data Management and Analysis Methods Beijing China Guangzhou China Department of Computer Science & Engineering The Hong Kong University of Science and Technology Hong Kong

Contrastive learning (CL)-based self-supervised learning models learn visual representations in a pairwise manner. Although the prevailing CL model has achieved great progress, in this paper, we uncover an ever-overlooked phenomenon: When the CL model is trained with full images, the performance tested in full images is better than that in foreground areas;when the CL model is trained with foreground areas, the performance tested in full images is worse than that in foreground areas. This observation reveals that backgrounds in images may interfere with the model learning semantic information and their influence has not been fully eliminated. To tackle this issue, we build a Structural Causal Model (SCM) to model the background as a confounder. We propose a backdoor adjustment-based regularization method, namely Interventional Contrastive Learning with Meta Semantic Regularizer (ICL-MSR), to perform causal intervention towards the proposed SCM. ICL-MSR can be incorporated into any existing CL methods to alleviate background distractions from representation learning. Theoretically, we prove that ICL-MSR achieves a tighter error bound. Empirically, our experiments on multiple benchmark datasets demonstrate that ICL-MSR is able to improve the performances of different state-of-the-art CL methods. Copyright © 2022, The Authors. All rights reserved.

关键词： Semantics

QoE Models for Online Video Streaming

学校读者我要写书评

暂无评论

QoE Models for Online Video Streaming

International Conference on Ubiquitous Computing and Communications and International Symposium on Cyberspace and Security (IUCC-CSS)

作者： Rongqi Zhang Xu Zhang Peiyao Guo Qilin Fan Hao Yin Zhan Ma School of Electronic Science and Engineering Nanjing University China Mathematics & Physical Sciences College of Engineering University of Exeter UK School of Big Data and Software Engineering Chongqing University China Chongqing Key Laboratory of Digital Cinema Art Theory and Technology Chongqing University China Beijing National Research Center for Information Science and Technology Tsinghua University China

ISBN: (纸本)9781665466684

With the rising popularity of video streaming services, Quality of Experience (QoE) acts as the crucial role in improving user's experience. Distinguished from Quality of Service (QoS), which mainly weigh the video streaming quality with network transmission performance, QoE focuses on the subjective and user-oriented assessment when users views the videos online. It is a vital issue for online streaming to ensure a user-satisfied transmission under dynamic network. This highlights the need for an accurate and feasible QoE model to balance the user experience and available transmission bandwidth. This paper summaries and analyzes existing explorations on QoE models for online video streaming, especially with the advancement on emerged learning-based models, to promote the development of this research field.

关键词： Analytical models Computational modeling Quality of service Bandwidth Streaming media Ubiquitous computing User experience

PRAD: Unsupervised KPI Anomaly Detection by Joint Prediction and Reconstruction of Multivariate Time Series

学校读者我要写书评

暂无评论

PRAD: Unsupervised KPI Anomaly Detection by Joint Prediction...

Symposia and Workshops on Ubiquitous, Autonomic and Trusted Computing, UIC-ATC

作者： Zhiying Xiong Qilin Fan Kai Wang Xiuhua Li Xu Zhang Qingyu Xiong School of Big Data and Software Engineering Chongqing University Chongqing China Chongqing Key Laboratory of Digital Cinema Art Theory and Technology Chongqing University Chongqing China School of Computer Science and Technology Harbin Institute of Technology Weihai China Research Institute of Cyberspace Security Harbin Institute of Technology Weihai China Haihe Laboratory of Information Technology Application Innovation Tianjin China Department of Computer Science University of Exeter Exeter UK

Detecting anomalies for key performance indicator (KPI) data is of paramount importance to ensure the quality and reliability of network services. However, building the anomaly detection system for KPI is challenging due to the scarcity of abnormal labels and highly dynamic or even unseen patterns. In this paper, we propose an unsupervised KPI anomaly detection framework named PRAD by jointly optimizing prediction-based and reconstruction-based modules. Specifically, PRAD employs two parallel graph attention networks (GATs) to learn metric-oriented and time-oriented relationships. A temporal convolutional attention network (TCAN) is used to capture complex long-term dependencies. Furthermore, to tackle the overfitting issue in the variational auto-encoder (VAE), PRAD designs a reconstruction-based module that combines VAE with a generative adversarial network (GAN). The experimental results on the two public datasets demonstrate that the proposed PRAD outperforms existing anomaly detection methods.

关键词：

Foster noisy label learning by exploiting noise-induced distortion in foreground localization

学校读者我要写书评

暂无评论

Neural Networks 2025年

作者： Ang Chen Feng Xu Xin Lyu Tao Zeng Xin Li College of Computer Science and Software Engineering Hohai University Nanjing 211100 China School of Computer Engineering Jiangsu Ocean University Lianyungang 222005 China Key Laboratory of Water Big Data Technology of Ministry of Water Resources Hohai University Nanjing 211100 China

Large-scale, well-annotated datasets are crucial for training deep neural networks. However, the prevalence of noisy-labeled samples can cause irreversible impairment to the generalization of models. Existing approaches have attempted to mitigate the impact of noisy labels by exploiting the different loss or confidence distributions between clean and noisy data to detect and correct noisy labels. This paper investigates the noise-induced distorting effect on foreground localization by tracking the model’s spatial attention distribution on visual activation maps. We observe that in clean samples, highly responsive regions usually focus on label-relevant foreground regions, whereas in noisy data, the model accidentally attends to uninformative background regions or cluttered object edges due to interference from label noise. Inspired by the observations, we propose a novel two-stage foreground localization-augmented noisy label learning framework named FLSC to concurrently boost the accuracy of sample selection and label correction for robust training. Specifically, FLSC first quantifies noise-induced distortion in foreground localization to foster conventional loss-based selection criteria by calculating the information reduction when deriving the foreground images from the original images based on the attention distribution. Next, we propose a noise-adaptive adversarial erasing strategy, which suppresses background activation by imposing adaptive erasure regularization, to eliminate overfitting to noisy samples while enhancing the learning of robust representations. To the best of our knowledge, it’s the first attempt to exploit localization quality evaluation based on feature activation to address the label noise problem. Extensive experiments on synthetic and real-world datasets validate the superior performance of FLSC compared to state-of-the-art methods.

关键词：

Sample hardness based gradient loss for long-tailed cervical cell detection

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Liu, Minmin Li, Xuechen Gao, Xiangbo Chen, Junliang Shen, Linlin Wu, Huisi Computer Vision Institute School of Computer Science and Software Engineering Shenzhen University China Shenzhen Institute of Artificial Intelligence of Robotics of Society Shenzhen China Guangdong Key Laboratory of Intelligent Information Processing Shenzhen University Shenzhen518060 China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen518060 China University of California Irvine United States

Due to the difficulty of cancer samples collection and annotation, cervical cancer datasets usually exhibit a long-tailed data distribution. When training a detector to detect the cancer cells in a WSI (Whole Slice Image) image captured from the TCT (Thinprep Cytology Test) specimen, head categories (e.g. normal cells and inflammatory cells) typically have a much larger number of samples than tail categories (e.g. cancer cells). Most existing state-of-the-art long-tailed learning methods in object detection focus on category distribution statistics to solve the problem in the long-tailed scenario, without considering the "hardness" of each sample. To address this problem, in this work we propose a Grad-Libra Loss that leverages the gradients to dynamically calibrate the degree of hardness of each sample for different categories, and re-balance the gradients of positive and negative samples. Our loss can thus help the detector to put more emphasis on those hard samples in both head and tail categories. Extensive experiments on a long-tailed TCT WSI image dataset show that the mainstream detectors, e.g. RepPoints, FCOS, ATSS, YOLOF, etc. trained using our proposed Gradient-Libra Loss, achieved much higher (7.8%) mAP than that trained using cross-entropy classification loss. Copyright © 2022, The Authors. All rights reserved.

关键词： Cytology

Detail retention and enhancement for camouflaged object detection

学校读者我要写书评

暂无评论

Neurocomputing 2025年 648卷

作者： Shangjing Chen Feng Xu Tao Zeng Xin Li Jiale Chen Xin Lyu College of Computer Science and Software Engineering Hohai University Nanjing 211100 China School of Computer Engineering Jiangsu Ocean University Lianyungang 222005 China Key Laboratory of Water Big Data Technology of Ministry of Water Resources Hohai University Nanjing 211100 China

The objective of camouflaged object detection is to segment objects that share similar visual attributes with their backgrounds. Existing methods leverage the consistency and variability of the object to pinpoint its location; however, when presented with samples exhibiting intricate edges, the model tends to overlook critical details. Toward this issue, we analyzed the reasons from the perspective of processing basic features and found that roughly compressing the feature channel could lead to the loss of fine features at the initial stage of the pipeline. Therefore, we introduce a multi-scale feature preprocessing approach named Soft Channel Compression prior to feature fusion, which retains key cues by enhancing the visibility of fine-grained features. We then utilize self-attention to fuse the results of multi-scale feature sampling, synthesizing contextual information between extended details and the main object. Furthermore, we propose the Hierarchical Receptive Field Complementarity module, using a divide-and-conquer strategy to enhance discriminative features with different scales in object details. This facilitates the integration of information across various ranges by employing a hierarchical manner to expand the receptive field, thereby achieving complementary information between local areas. Extensive experimental results on benchmark datasets demonstrate that the proposed method outperforms other state-of-the-art methods.

关键词： Camouflaged object detection Edge detail Hierarchical manner Multi-scale feature Regional information complementarity