检索结果-内蒙古大学图书馆

2023 IEEE International Conference on Big data, Bigdata 2023

作者： Flores, Ricardo Shrestha, Avantika Rundensteiner, Elke A. Worcester Polytechnic Institute Data Science WorcesterMA United States Worcester Polytechnic Institute Computer Science/Data Science WorcesterMA United States

ISBN: (纸本)9798350324457

Depression is a prevalent mental health condition with severe impacts on physical and social health. It is costly and difficult to detect, requiring substantial time from trained mental professionals. To alleviate this burden, recent research explores the diagnostic capabilities of deep learning models trained on modalities extracted from videos of clinical interviews for depression screening. However, training deep learning models is challenging because in the mental health domain datasets contain a small number of patients. To address this challenge, we propose DeepScreen, a recurrent deep-learning architecture for depression screening whose performance is boosted by deploying a self-supervised auxiliary task for selective missing value imputation. DeepScreen leverages a multi-task architecture with a bidirectional recurrent deep learning model and a self-attention mechanism, jointly optimizing the supervised depression and the self-supervised auxiliary task. Our first study assesses the capability of the auxiliary task training of DeepScreen under different correlation levels and masking sub-sequence sizes of multi-variate time series. Found to be effective, our second study evaluates DeepScreen on 15 data sets composed of real-world temporal facial landmark features extracted from responses to different clinical interview questions. The results achieved across all 15 datasets demonstrate that the imputation task boosts the depression prediction metrics significantly. Specifically, DeepScreen improves the F1 score for one of the datasets by 57%. Further, our best-performing model achieves an F1 score of 0.85. This work provides valuable insights into improving deep learning driven mental screening applications by leveraging auxiliary tasks such as imputation for learning a better representation even from small datasets. © 2023 IEEE.

关键词： Digital health missing value imputation multivariate time series sequential modeling for mental health screen

来源：评论

学校读者我要写书评

暂无评论

An Effective Dynamic Gradient Calibration Method for Continual Learning 41

An Effective Dynamic Gradient Calibration Method for Continu...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Lin, Weichen Chen, Jiaxiang Huang, Ruomin Ding, Hu School of Data Science University of Science and Technology of China Anhui China Duke University United States School of Computer Science and Technology University of Science and Technology of China Anhui China

Continual learning (CL) is a fundamental topic in machine learning, where the goal is to train a model with continuously incoming data and tasks. Due to the memory limit, we cannot store all the historical data, and therefore confront the "catastrophic forgetting" problem, i.e., the performance on the previous tasks can substantially decrease because of the missing information in the latter period. Though a number of elegant methods have been proposed, the catastrophic forgetting phenomenon still cannot be well avoided in practice. In this paper, we study the problem from the gradient perspective, where our aim is to develop an effective algorithm to calibrate the gradient in each updating step of the model;namely, our goal is to guide the model to be updated in the right direction under the situation that a large amount of historical data are unavailable. Our idea is partly inspired by the seminal stochastic variance reduction methods (e.g., SVRG and SAGA) for reducing the variance of gradient estimation in stochastic gradient descent algorithms. Another benefit is that our approach can be used as a general tool, which is able to be incorporated with several existing popular CL methods to achieve better performance. We also conduct a set of experiments on several benchmark datasets to evaluate the performance in practice. Copyright 2024 by the author(s)

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Dual VC Dimension Obstructs Sample Compression by Embeddings 37

Dual VC Dimension Obstructs Sample Compression by Embeddings

引用

37th Annual Conference on Learning Theory, COLT 2024

作者： Chase, Zachary Chornomaz, Bogdan Hanneke, Steve Moran, Shay Yehudayoff, Amir Department of Mathematics Technion Israel Department of Computer Science Purdue University United States Departments of Mathematics Computer Science and Data and Decision Sciences Technion Google Research Israel Department of Computer Science Copenhagen University Denmark

This work studies embedding of arbitrary VC classes in well-behaved VC classes, focusing particularly on extremal classes. Our main result expresses an impossibility: such embeddings necessarily require a significant increase in dimension. In particular, we prove that for every d there is a class with VC dimension d that cannot be embedded in any extremal class of VC dimension smaller than exponential in d. In addition to its independent interest, this result has an important implication in learning theory, as it reveals a fundamental limitation of one of the most extensively studied approaches to tackling the long-standing sample compression conjecture (Warmuth, 2003). Concretely, the approach proposed by Floyd and Warmuth (1995) entails embedding any given VC class into an extremal class of a comparable dimension, and then applying an optimal sample compression scheme for extremal classes. However, our results imply that this strategy would in some cases result in a sample compression scheme at least exponentially larger than what is predicted by the sample compression conjecture. The above implications follow from a general result we prove: any extremal class with VC dimension d has dual VC dimension at most 2d + 1. This bound is exponentially smaller than the classical bound 2d+1-1 of Assouad (1983), which applies to general concept classes (and is known to be unimprovable for some classes). We in fact prove a stronger result, establishing that 2d + 1 upper bounds the dual Radon number of extremal classes. This theorem represents an abstraction of the classical Radon theorem for convex sets, extending its applicability to a wider combinatorial framework, without relying on the specifics of Euclidean convexity. The proof utilizes the topological method and is primarily based on variants of the Topological Radon Theorem (Bajmóczy and Bárány, 1979). © 2024 Z. Chase, B. Chornomaz, S. Hanneke, S. Moran & A. Yehudayoff.

关键词： Topology

来源：评论

学校读者我要写书评

暂无评论

An Architecture for Dynamic Load Balancing in Cloud Environment 2

An Architecture for Dynamic Load Balancing in Cloud Environm...

引用

2nd International Conference on Edge Computing and Applications, ICECAA 2023

作者： Lokesh, Gudivada Baseer, K.K. Dept. of Computer Science and Engineering Tirupati India Department of Data Science Tirupati India

ISBN: (纸本)9798350347579

Clouds are highly customizable infrastructures that offer a platform as a service and let customers subscribe on a pay-as-you-go basis to their requirements. The straightforward service-oriented cloud computing model is gaining popularity around the world. The number of people using the Cloud is constantly growing. Clouds use modern data centers to manage a massive number of users. The reliability of the Cloud depends on load balancing. Balancing virtual machine loads lowers energy consumption and task rejections by optimizing resource utilization. One can increase performance while using fewer resources using load balancing, resource management, quality of service, etc. The difficulty of overloading and underloading virtual machines in cloud computing can be lessened by load balancing in the *** research study thoroughly examined the load-balancing algorithms found in the literature. First, the traditional approaches are analyzed before moving on to more recent work on load balancing with heterogeneous techniques. Along with the tools available for the current investigation, various metrics are used to evaluate the load-balancing algorithms. The proposed article will primarily serve to assist in the development of new algorithms in the future. © 2023 IEEE.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

An Innovative K-Anonymity Privacy-Preserving Algorithm to Improve data Availability in the Context of Big data

引用

computers, Materials & Continua 2024年第4期79卷 1561-1579页

作者： Linlin Yuan Tiantian Zhang Yuling Chen Yuxiang Yang Huang Li State Key Laboratory of Public Big Data College of Computer Science and TechnologyGuizhou UniversityGuiyang550025China College of Information Engineering Guizhou Open UniversityGuiyang550025China Guizhou Academy of Tobacco Science Guiyang550025China

The development of technologies such as big data and blockchain has brought convenience to life,but at the same time,privacy and security issues are becoming more and more *** K-anonymity algorithm is an effective and low computational complexity privacy-preserving algorithm that can safeguard users’privacy by anonymizing big ***,the algorithm currently suffers from the problem of focusing only on improving user privacy while ignoring data *** addition,ignoring the impact of quasi-identified attributes on sensitive attributes causes the usability of the processed data on statistical analysis to be *** on this,we propose a new K-anonymity algorithm to solve the privacy security problem in the context of big data,while guaranteeing improved data ***,we construct a new information loss function based on the information quantity *** that different quasi-identification attributes have different impacts on sensitive attributes,we set weights for each quasi-identification attribute when designing the information loss *** addition,to reduce information loss,we improve K-anonymity in two ***,we make the loss of information smaller than in the original table while guaranteeing privacy based on common artificial intelligence algorithms,i.e.,greedy algorithm and 2-means clustering *** addition,we improve the 2-means clustering algorithm by designing a mean-center method to select the initial center of ***,we design the K-anonymity algorithm of this scheme based on the constructed information loss function,the improved 2-means clustering algorithm,and the greedy algorithm,which reduces the information ***,we experimentally demonstrate the effectiveness of the algorithm in improving the effect of 2-means clustering and reducing information loss.

关键词： Blockchain big data K-anonymity 2-means clustering greedy algorithm mean-center method

来源：评论

学校读者我要写书评

暂无评论

N-of-one differential gene expression without control samples using a deep generative model

引用

Genome Biology 2023年第1期24卷 1页

作者： Prada-Luengo, Iñigo Schuster, Viktoria Liang, Yuhu Terkelsen, Thilde Sora, Valentina Krogh, Anders Department of Computer Science University of Copenhagen Copenhagen Denmark Center for Health Data Science University of Copenhagen Copenhagen Denmark

Differential analysis of bulk RNA-seq data often suffers from lack of good controls. Here, we present a generative model that replaces controls, trained solely on healthy tissues. The unsupervised model learns a low-dimensional representation and can identify the closest normal representation for a given disease sample. This enables control-free, single-sample differential expression analysis. In breast cancer, we demonstrate how our approach selects marker genes and outperforms a state-of-the-art method. Furthermore, significant genes identified by the model are enriched in driver genes across cancers. Our results show that the in silico closest normal provides a more favorable comparison than control samples. © 2023, The Author(s).

关键词： Deep generative models Deep learning DEG DEseq2 Differential expression analysis Transcriptomics

来源：评论

学校读者我要写书评

暂无评论

PDNet: A Novel Approach to Plant Disease Detection using CNN and Keras 6

PDNet: A Novel Approach to Plant Disease Detection using CNN...

引用

6th International Conference on Mobile Computing and Sustainable Informatics, ICMCSI 2025

作者： Jain, Garima Shukla, Amita Chaudhary, Anamika Sharma, Nidhi Bisht, Sovers Singh Jain, Ankush Department of Computer Science and Buisness Systems Noida Institute of Engineering and Technology Greater Noida India Department of Data Science Noida Institute of Engineering and Technology Greater Noida India Department of Data Science Noida Institute of Engineering and Technology Delhi India Department of Computer Science and Engineering Netaji Subhas University of Technology Delhi India

ISBN: (纸本)9798331522667

This research addresses the pressing global demand for food by leveraging cutting-edge deep learning techniques for automating plant disease detection. Focusing on tomato and potato leaf diseases, the study utilized the PlantVillage dataset to compare classical machine learning methods, such as Local Binary Patterns (LBP) combined with classifiers like kNN, SVM, and Extreme Learning Machines, against modern Convolutional Neural Network (CNN) architectures. These included MobileNet, DenseNet201, ResNet50, Inception V3, and VGG16/19. A custom-designed CNN model was developed with unique configurations implemented using Keras and TensorFlow. The model was tested for binary, 6-class, and 10-class classifications, achieving remarkable accuracy rates of 99.6%, 98.4%, and 97.2%, respectively. Notably, after 50 training epochs, the model achieved a peak accuracy of 99.78% for plant disease classification. A practical web-based application was also developed, integrating the best-performing model to facilitate real-time leaf disease detection and notify farmers via email. This approach demonstrates a significant advancement in sustainable agriculture by providing an efficient, accurate, and scalable solution for early disease detection, which is critical for improving crop yield and productivity. © 2025 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Tree-Based Genetic Programming for Evolutionary Analog Circuit with Approximate Shapley Value 44th

Tree-Based Genetic Programming for Evolutionary Analog Circ...

引用

44th SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence, SGAI 2024

作者： Shi, Xinming Minku, Leandro L. Yao, Xin School of Electronics Electrical Engineering and Computer Science Queen’s University Belfast Belfast United Kingdom School of Computer Science University of Birmingham Birmingham United Kingdom School of Data Science Lingnan University China

ISBN: (纸本)9783031779145

The automated design of analog circuits presents a significant challenge due to the complexity of circuit topology and parameter selection. Traditional evolutionary algorithms, such as Genetic Programming (GP), have shown potential in this domain but are often hindered by inefficient search processes and the large design space. Furthermore, fitness evaluation in the evolutionary design of circuits is often computationally very expensive. In this paper, we introduce a novel evolutionary framework that leverages approximate Shapley values to guide the optimization process in tree-based genetic programming for analog circuit design. Our approach addresses the computational challenges associated with computing Shapley values by introducing a two-stage evolutionary framework that includes a Shapley Value Library (SVlib) and a KNN-based prediction for efficient estimation of Shapley values. Our proposed work not only enhances the search efficiency by focusing on the most beneficial sub-circuits but also leads to more compact and efficient circuit designs. Furthermore, fitness evaluation in the evolutionary design of circuits is often computationally very expensive experiments, we verify that our framework accelerates evolutionary convergence and outperforms traditional methods in terms of circuit optimization. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Genetic programming

来源：评论

学校读者我要写书评

暂无评论

Sequentia12D: Organizing Center of Skip Connections for Transformers 22

Sequentia12D: Organizing Center of Skip Connections for Tran...

引用

22nd IEEE International Conference on Machine Learning and Applications, ICMLA 2023

作者： Pathak, Harsh Nilesh Paffenroth, Randy Hershey, Quincy Data Science Worcester Polytechnic Institute Worcester United States Mathematical Sciences Computer Science and Data Science Worcester Polytechnic Institute Worcester United States

ISBN: (纸本)9798350345346

Skip connections (SCs) are commonly employed in neural networks to facilitate gradient-based training and often lead to improved performance in deep learning. To implement SCs, a user writes custom modules along with *** which limits the extensibility of neural network design with SC. Despite their versatility and user-friendliness, there is an opportunity to create a standardized approach for organizing SCs in various applications. Establishing a clear organizing framework would significantly enhance the widespread adoption of this powerful technique, amplifying its impact. We introduce Sequential2D as an enhanced orchestrator which provides in-sights into not only the width and depth of the network but also a neural network's intra-connections such as SCs. The functionality provided by the Sequential2D architecture is a strict super-set of the functionality found in the common *** network. The 2D in Sequential2D refers to the two-dimensional matrix of functions, as opposed to the one-dimensional cascading array of functions in a Pytorch or Keras sequential container. As an example of our proposed approach, we demonstrate the effectiveness of our Sequential2D architecture by substantially improving the fine-tuning performance of the GPT-2 model with less than 1% additional parameters. © 2023 IEEE.

关键词： deep learning GPT-2 Neural Network nlp recurrent neural networks RNN Sequential skip-connection sparsity transformer

来源：评论

学校读者我要写书评

暂无评论

Adaptive Feature Selection on Graph 9

Adaptive Feature Selection on Graph

引用

9th International Conference on Information, Cybernetics, and Computational Social Systems, ICCSS 2023

作者： Lin, Yue-Na Zhang, Chun-Yang School of Computer and Data Science Fuzhou University Fuzhou China

ISBN: (纸本)9798350342239

Graphs have become a widely-used tool to model data with relationships in real life for a long time. To discover the important contents in the graph, many graph neural networks (GNNs) have been come up with. Nevertheless, these models tend to adopt ReLU as their activation function for its nonlinearity, effectiveness, and efficiency. Owing to its own deficiency, it would cause many generated feature elements to be zero, which would miss part significant features. To overcome this problem, an adaptive weight vector to tune the features was provided. By limiting the elements of the weight vector, it can be a better substitute for ReLU. Besides, such a weight vector adaptively measures the importance of each feature element to work as a feature selection operator. To show the function of the weight vector, we examine a GCN with the weight vecotor in node classification, and it exhibits overall improvements over three well-known citation networks. © 2023 IEEE.

关键词： Feature Selection

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：