检索结果-内蒙古大学图书馆

Plant Disease Identification under Imbalanced dataset using Hybrid Deep Learning Method

Journal of The Institution of Engineers (India): Series A 2025年第1期106卷 19-29页

作者： Zhou, Changjian Zhang, Xin Department of Data and Computing Northeast Agricultural University Harbin150030 China School of Electrical and Information Northeast Agricultural University Harbin150030 China

Recent studies suggest that plant disease identification via computational approaches is vital for agricultural production. However, there is still a large gap that needs to be bridged while the training data is imbalanced when the number of samples in different categories of the dataset varies greatly. To solve this limitation, a hybrid deep learning method combining deep residual network, dense network, and deep convolution generative adversarial network (DCGAN) is proposed for plant disease identification in this work, which takes advantage of these three models. Including 34,501 original images with 33 categories are collected, where one category contains 4442 samples and another contains 74, causing the phenomenon of data imbalance. Importantly, the imbalanced dataset has a negative impact on training performance. To address this issue, the DCGAN is introduced for data augmentation to make up for the limit of training data in this research. In addition, the residual and dense network are combined as a novel deep learning method to improve prediction ability. Together, the original and generated images are integrated as a mixed dataset for training, and only the original images were utilized for testing. Experimental results indicated that the presented approach achieved 0.977 F1-score and 0.987 test accuracy, outperformed the existing state-of-the-art models. These findings indicate that the hybrid deep learning approach, through the ingenious integration of the strengths of three sub-networks, significantly enhances the generalization capability of the model. This methodology not only optimizes overall performance but also underscores the profound potential in tackling these complex problems. Furthermore, a smartphone-based point-to-point identification system was designed to provide convenience for users in practical application. © The Institution of Engineers (India) 2024.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

The Implicit Bias of Adam on Separable data 38

The Implicit Bias of Adam on Separable Data

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Zhang, Chenyang Zou, Difan Cao, Yuan Department of Statistics and Actuarial Science School of Computing and Data Science The University of Hong Kong Hong Kong Department of Computer Science School of Computing and Data Science Institute of Data Science The University of Hong Kong Hong Kong Department of Statistics and Actuarial Science School of Computing and Data Science Department of Mathematics The University of Hong Kong Hong Kong

Adam has become one of the most favored optimizers in deep learning problems. Despite its success in practice, numerous mysteries persist regarding its theoretical understanding. In this paper, we study the implicit bias of Adam in linear logistic regression. Specifically, we show that when the training data are linearly separable, the iterates of Adam converge towards a linear classifier that achieves the maximum ∞-margin in direction. Notably, for a general class of diminishing learning rates, this convergence occurs within polynomial time. Our result shed light on the difference between Adam and (stochastic) gradient descent from a theoretical perspective. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ARCHER:a ReRAM-based accelerator for compressed recommendation systems

引用

Frontiers of Computer Science 2024年第5期18卷 147-160页

作者： Xinyang SHEN Xiaofei LIAO Long ZHENG Yu HUANG Dan CHEN Hai JIN National Engineering Research Center for Big Data Technology and System Services Computing Technology and System LabClusters and Grid Computing LabSchool of Computer Science and TechnologyHuazhong University of Science and TechnologyWuhan 430074China

Modern recommendation systems are widely used in modern data *** random and sparse embedding lookup operations are the main performance bottleneck for processing recommendation systems on traditional platforms as they induce abundant data movements between computing units and ***-based processing-in-memory(PIM)can resolve this problem by processing embedding vectors where they are ***,the embedding table can easily exceed the capacity limit of a monolithic ReRAM-based PIM chip,which induces off-chip accesses that may offset the PIM ***,we deploy the decomposed model on-chip and leverage the high computing efficiency of ReRAM to compensate for the decompression performance *** this paper,we propose ARCHER,a ReRAM-based PIM architecture that implements fully yon-chip recommendations under resource ***,we make a full analysis of the computation pattern and access pattern on the decomposed *** on the computation pattern,we unify the operations of each layer of the decomposed model in multiply-and-accumulate *** on the access observation,we propose a hierarchical mapping schema and a specialized hardware design to maximize resource *** the unified computation and mapping strategy,we can coordinatethe inter-processing elements *** evaluation shows that ARCHER outperforms the state-of-the-art GPU-based DLRM system,the state-of-the-art near-memory processing recommendation system RecNMP,and the ReRAM-based recommendation accelerator REREC by 15.79×,2.21×,and 1.21× in terms of performance and 56.06×,6.45×,and 1.71× in terms of energy savings,respectively.

关键词： recommendation system ReRAM processing-in-memory embedding layer

来源：评论

学校读者我要写书评

暂无评论

FPS Killer - A Moving Target Detection Algorithm in FPS Game Based on Yolov3 Frame and Gradient Descent Algorithm 2

FPS Killer - A Moving Target Detection Algorithm in FPS Game...

引用

2nd International Conference on Advancements in Smart, Secure and Intelligent computing, ASSIC 2024

作者： Huang, Jiayu Liu, Aofan School of Computing Jiangsu Vocational College of Commerce Jiangsu China School of Computing & Data Science Xiamen University Malaysia Sepang Malaysia

ISBN: (纸本)9798350370188

With the continuous development of the economic level of the society, the game industry is getting more and more extensive development among young people. As one of the mainstreams in the game industry, FPS games are warmly welcomed by the majority of players. In order to ensure the fairness of the game, anti-cheat technologies have emerged, such as 'BattlEye' and 'ValveAnti-Cheat'. This article applies the current popular computer vision technology to FPS games and proposes a cheating technology that cannot be detected by traditional anti-cheating methods. However, this project will only be used for academic research and discussion, and it is forbidden to use it for commercial sales and esports. © 2024 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Two genetic algorithms for final exam scheduling 24

Two genetic algorithms for final exam scheduling

引用

6th Asia Conference on Machine Learning and computing, ACMLC 2024

作者： Mixer, Mark Morrow, Steven Sukherman, Lev School of Computing and Data Science Wentworth Institute of Technology BostonMA United States

ISBN: (纸本)9798400710018

The scheduling of final exams at a university is a problem which can be improved with artificial intelligence techniques. In this paper we explain and compare two algorithms used to solve the exam scheduling problem at Wentworth Institute of Technology. The first technique is an implementation of a pure genetic algorithm, while the second combines some graph theoretical algorithms with two variations of a genetic algorithm. We will compare the performance of these approaches to each other and to historical results. © 2024 Copyright held by the owner/author(s).

关键词： Scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Uncertainty-Calibrated Test-Time Model Adaptation without Forgetting

引用

IEEE Transactions on Pattern Analysis and Machine Intelligence 2025年 PP卷 1-14页

作者： Tan, Mingkui Chen, Guohao Wu, Jiaxiang Zhang, Yifan Chen, Yaofo Zhao, Peilin Niu, Shuaicheng South China University of Technology School of Software Engineering China Pazhou Laboratory Guangzhou China XVERSE China National University of Singapore School of Computing Singapore Tencent AI Lab China Nanyang Technological University College of Computing and Data Science Singapore

Test-time adaptation (TTA) seeks to tackle potential distribution shifts between training and testing data by adapting a given model w.r.t. any testing sample. This task is particularly important when the test environment changes frequently. Although some recent attempts have been made to handle this task, we still face two key challenges: 1) prior methods have to perform backpropagation for each test sample, resulting in unbearable optimization costs to many applications;2) while existing TTA solutions can significantly improve the test performance on out-of-distribution data, they often suffer from severe performance degradation on in-distribution data after TTA (known as catastrophic forgetting). To this end, we have proposed an Efficient Anti-Forgetting Test-Time Adaptation (EATA) method which develops an active sample selection criterion to identify reliable and non-redundant samples for test-time entropy minimization. To alleviate forgetting, EATA introduces a Fisher regularizer estimated from test samples to constrain important model parameters from drastic changes. However, in EATA, the adopted entropy loss consistently assigns higher confidence to predictions even when the samples are underlying uncertain, leading to overconfident predictions that underestimate the data uncertainty. To tackle this, we further propose EATA with Calibration (EATA-C) to separately exploit the reducible model uncertainty and the inherent data uncertainty for calibrated TTA. Specifically, we compare the divergence between predictions from the full network and its sub-networks to measure the reducible model uncertainty, on which we propose a test-time uncertainty reduction strategy with divergence minimization loss to encourage consistent predictions instead of overconfident ones. To further re-calibrate predicting confidence on different samples, we utilize the disagreement among predicted labels as an indicator of the data uncertainty. Based on this, we devise a min-max entropy

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Patterns Recognition of Sentimental Patterns from the X-dataset using Deep Learning Algorithm 4

Patterns Recognition of Sentimental Patterns from the X-Data...

引用

4th International Conference on Soft computing for Security Applications, ICSCSA 2024

作者： Konduri, Praveen S R Department of Data Science School of Computing Mohan Babu University Tirupati India

ISBN: (纸本)9798331515720

This research investigates the application of deep learning techniques to sentiment analysis, a field focused on mining online platforms for subjective assessments. By combining deep neural networks (DNNs) with particle swarm optimization (PSO), we propose a hybrid method to improve feature selection and enhance the accuracy of sentiment classification. The study employs a dataset collected from Twitter using the Ruby Twitter API and evaluates the performance of the proposed method using k-fold cross-validation. The results are compared with the previously used firefly search (FS) method to demonstrate the effectiveness of the hybrid approach. This research contributes to the advancement of sentiment analysis techniques and their application in various domains, such as social media monitoring and market research. © 2024 IEEE.

关键词： Particle swarm optimization (PSO)

来源：评论

学校读者我要写书评

暂无评论

Cube Attacks on Round-Reduced Grain-128AEAD 6th

Cube Attacks on Round-Reduced Grain-128AEAD

引用

6th International Conference on Emerging Technologies in computing, iCETiC 2023

作者： Teng, Wil Liam Salam, Iftekhar Yau, Wei-Chuen Teh, Jia Yew School of Computing and Data Science Xiamen University Malaysia Sepang43900 Malaysia

ISBN: (纸本)9783031502149

Lightweight cryptography aims to design secure and efficient cryptographic algorithms for resource-constrained devices. Traditional cryptographic algorithms may not be readily usable in resource-constrained environments. To standardise cryptographic solutions tailored for such resource-constraint environments, the National Institute of Standards and Technology (NIST) launched the Lightweight Cryptography (LWC) project. Grain-128AEAD is a stream cipher-based finalist in the NIST LWC project. In this work, we examine the security of the initial version of Grain-128AEAD against cube attacks. We present distinguishing attacks on a reduced-round version of the cipher, assuming that the keystream can be observed immediately after the reduced-round initialisation of the pre-output generator. We obtained various cubes of sizes 25 to 45 for reduced-round Grain-128AEAD. The best cube reported in this work can distinguish the output of a 165-round initialisation of Grain-128AEAD with a cube size of 35. The complexity of the distinguishing attack is O(235). The results are confirmed experimentally. We conclude that even with fewer rounds of initialisation for the first version of Grain-128AEAD, the cipher still has a good security margin against cube attacks. © 2024, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.

关键词： Geometry

来源：评论

学校读者我要写书评

暂无评论

Bayesian Inverse Transfer in Evolutionary Multiobjective Optimization

引用

ACM Transactions on Evolutionary Learning and Optimization 2024年第4期4卷 1-27页

作者： Liu, Jiao Gupta, Abhishek Ong, Yew-Soon College of Computing and Data Science Nanyang Technological University Singapore Singapore School of Mechanical Sciences Indian Institute of Technology Goa Goa India

Transfer optimization enables data-efficient optimization of a target task by leveraging experiential priors from related source tasks. This is especially useful in multiobjective optimization settings where a set of tradeoff solutions is sought under tight evaluation budgets. In this article, we introduce a novel concept of inverse transfer in multiobjective optimization. Inverse transfer stands out by employing Bayesian inverse Gaussian process models to map performance vectors in the objective space to population search distributions in task-specific decision space, facilitating knowledge transfer through objective space unification. Building upon this idea, we introduce the first Inverse Transfer Evolutionary Multiobjective Optimizer (invTrEMO). A key highlight of invTrEMO is its ability to harness the common objective functions prevalent in many application areas, even when decision spaces do not precisely align between tasks. This allows invTrEMO to uniquely and effectively utilize information from heterogeneous source tasks as well. Furthermore, invTrEMO yields high-precision inverse models as a significant byproduct, enabling the generation of tailored solutions on-demand based on user preferences. Empirical studies on multi-and many-objective benchmark problems, as well as a practical case study, showcase the faster convergence rate and modeling accuracy of the invTrEMO relative to state-of-The-Art evolutionary and Bayesian optimization algorithms. The source code of the invTrEMO is made available at https://***/LiuJ-2023/invTrEMO. © 2024 Copyright held by the owner/author(s).

关键词： Budget control

来源：评论

学校读者我要写书评

暂无评论

A survey on dynamic graph processing on GPUs: concepts, terminologies and systems

引用

Frontiers of Computer Science 2024年第4期18卷 1-23页

作者： Hongru GAO Xiaofei LIAO Zhiyuan SHAO Kexin LI Jiajie CHEN Hai JIN National Engineering Research Center for Big Data Technology and System/Services Computing Technology and System Lab/Cluster and Grid Computing Lab School of Computer Science and TechnologyHuazhong University of Science and TechnologyWuhan 430074China Zhejiang Lab Hangzhou 311121China

Graphs that are used to model real-world entities with vertices and relationships among entities with edges,have proven to be a powerful tool for describing real-world problems in *** most real-world scenarios,entities and their relationships are subject to constant *** that record such changes are called dynamic *** recent years,the widespread application scenarios of dynamic graphs have stimulated extensive research on dynamic graph processing systems that continuously ingest graph updates and produce up-to-date graph analytics *** the scale of dynamic graphs becomes larger,higher performance requirements are demanded to dynamic graph processing *** the massive parallel processing power and high memory bandwidth,GPUs become mainstream vehicles to accelerate dynamic graph processing ***-based dynamic graph processing systems mainly address two challenges:maintaining the graph data when updates occur(i.e.,graph updating)and producing analytics results in time(i.e.,graph computing).In this paper,we survey GPU-based dynamic graph processing systems and review their methods on addressing both graph updating and graph *** comprehensively discuss existing dynamic graph processing systems on GPUs,we first introduce the terminologies of dynamic graph processing and then develop a taxonomy to describe the methods employed for graph updating and graph *** addition,we discuss the challenges and future research directions of dynamic graph processing on GPUs.

关键词： dynamic graphs graph processing graph algorithms GPUs

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：