检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Lei, Shiye Zhang, Sen Tao, Dacheng School of Computer Science The University of Sydney Australia College of Computing & Data Science Nanyang Technological University Singapore

Massive reinforcement learning (RL) data are typically collected to train policies offline without the need for interactions, but the large data volume can cause training inefficiencies. To tackle this issue, we formulate offline behavior distillation (OBD), which synthesizes limited expert behavioral data from sub-optimal RL data, enabling rapid policy learning. We propose two naive OBD objectives, DBC and PBC, which measure distillation performance via the decision difference between policies trained on distilled data and either offline data or a near-expert policy. Due to intractable bi-level optimization, the OBD objective is difficult to minimize to small values, which deteriorates PBC by its distillation performance guarantee with quadratic discount complexity O(1/(1 − γ)2). We theoretically establish the equivalence between the policy performance and action-value weighted decision difference, and introduce action-value weighted PBC (Av-PBC) as a more effective OBD objective. By optimizing the weighted decision difference, Av-PBC achieves a superior distillation guarantee with linear discount complexity O(1/(1 − γ)). Extensive experiments on multiple D4RL datasets reveal that Av-PBC offers significant improvements in OBD performance, fast distillation convergence speed, and robust cross-architecture/optimizer generalization. The code is available at https://***/LeavesLei/OBD. © 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

Predicting Employee Churn- data Driven Approach for Retention Management

Predicting Employee Churn- Data Driven Approach for Retentio...

引用

2024 International Conference on System, Computation, Automation and Networking, ICSCAN 2024

作者： Jayaraj, R. Mehrotra, Safal Kumar, Harshit School of Computing Srm Institute of Science and Technology Dept. of Data Science and Business Systems Chennai India

ISBN: (纸本)9798331510022

Employee churn is a critical challenge faced by organizations across industries, leading to disruptions in productivity and increased costs associated with recruitment and training. In this study, we propose an effective solution leveraging Random Forests to predict employee churn accurately and efficiently. Our approach addresses the inherent complexities of employee turnover prediction while considering resource constraints commonly encountered in practical applications. The proposed model incorporates feature engineering techniques and ensemble learning principles to harness the predictive power of Random Forests. Through a comprehensive evaluation on real-world employee datasets, we demonstrate the superior performance of our model in comparison to existing state-of-the-art methods. Our analysis encompasses various evaluation metrics, including accuracy, precision, recall, and F1-score, showcasing the robustness and reliability of our model across diverse organizational contexts. © 2024 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Convolutional Neural Network-Based Multi-Fruit Classification and Quality Grading with a Gradio Interface

Convolutional Neural Network-Based Multi-Fruit Classificatio...

引用

2024 International Conference on Innovative computing, Intelligent Communication and Smart Electrical Systems, ICSES 2024

作者： Nandhini, E. Vadivu, G. School of Computing Srm Institute of Science and Technology Department of Data Science and Business Systems Kattankulathur Chennai India

ISBN: (数字)9798331543624

ISBN: (纸本)9798331543617

Agricultural products are experiencing significant losses, especially after the harvesting phase, which is the period from fruit storage to the product reaching the consumer's hands. A lack of human resources, time consumption, product handling, and human errors during the post-harvest stage of the fruit grading and sorting process result in huge losses. Deep learning (DL) offers an automated process for classifying and identifying fruits. Developing an architecture for a VGG16 convolutional neural network (CNN) within the Gradio user interface, utilizing five distinct fruit types as the model's dataset, such as apple, banana, guava, orange, apple, and pomegranate. For five different sets of fresh and defect fruits, the proposed model got 91.1% and 89.1% of test accuracy for multi-fruit clas- sification and fruit status prediction, respectively. This machine- based automated model has the potential to reduce losses and enhance fruit grading and sorting efficiency, thereby significantly impacting sustainable and productive agricultural industries through increased productivity. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Gesture-driven Innovation: Exploring the Intersection of Human-Computer Interaction and Virtual Fashion Try-On Systems 2

Gesture-driven Innovation: Exploring the Intersection of Hum...

引用

2nd IEEE International Conference on Networking and Communications, ICNWC 2024

作者： Prakash, M. Arunkumar, Nithes Gokul, G. SRM Instititue of Science and Technology School of Computing Department of Data Science and Business Systems Kattankulathur603203 India

ISBN: (纸本)9798350365269

Virtual Fashion Try-On systems revolutionize online fashion retail, offering immersive shopping experiences. This paper investigates integrating gesture-driven interaction into these systems to enhance engagement, realism, and usability. The paper highlights the importance of Virtual Fashion Try-On systems in addressing fit uncertainty and tactile interaction limitations in online shopping. It explores gesture-driven interaction to enrich virtual fashion experiences. The paper examines gesture recognition, covering hand gestures, body movements, and facial expressions. Various algorithms and techniques, including computer vision and machine learning, are discussed, addressing challenges like accuracy and real-time performance. Challenges and future directions in gesture-driven Virtual Fashion Try-On are discussed, covering technical hurdles, human-centered design, and ethical considerations. Recommendations for future research and development are outlined. © 2024 IEEE.

关键词： Virtual reality

来源：评论

学校读者我要写书评

暂无评论

An Intelligent Machine Learning Approach to Detect Spam in Social Media 5

An Intelligent Machine Learning Approach to Detect Spam in S...

引用

5th International Conference on data Intelligence and Cognitive Informatics, ICDICI 2024

作者： Mohandas, R. Prasanna, D.S. John Deva Meenakshi, N. Kaviyaraj, R. Nathiya, S. Arivazhagan, N. Srm Institute of Science and Technology School of Computing Department of Computational Intelligence Kattankulathur India Srm Institute of Science and Technology School of Computing Department of Data Science and Business Systems Kattankulathur India

ISBN: (纸本)9798350389609

Machine learning based intelligent approach is applied to finding spam in YouTube videos. Spam comments are ones that are promotional or unrelated. A growing number of users have been drawn to the idea of making money by placing advertisements on popular social networking platforms over time. These remarks damage a channel's reputation. YouTube has addressed the issue with a limited set of solutions that concentrate on preventing comments from including links. Such heuristics are now futile since spammers have found ways to get around them. Despite the fact that common machine learning classification techniques have demonstrated considerable success, new methods can still increase accuracy. Therefore, this research attempt to identify such statements using machine learning techniques like Random Forest, SVM, and Naive Bayes. In order to convey a clear concept, graphs are used to illustrate data cleaning, data analysis, text preparation, and model building in this research. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

End-to-end Compilation is All FPGAs Need: A Unified Overlay-based FPGA Compiler for Deep Learning 25

End-to-end Compilation is All FPGAs Need: A Unified Overlay-...

引用

30th Asia and South Pacific Design Automation Conference, ASP-DAC 2025

作者： Qian, Kai Lu, Haodong Liu, Yinqiu Zhang, Zexu Wang, Kun School of Microelectronics Fudan University China College of Computing and Data Science Nanyang Technological University Singapore

ISBN: (纸本)9798400706356

Field-Programmable Gate Array (FPGA) has shown great application potential in deploying Neural Networks (NNs) due to the characteristics of programmability, low power consumption, etc. However, deploying NNs on FPGA is non-trivial because (1) Mainstream NNs pose significant FPGA architecture design challenges due to their large number of parameters, complex operations, and the need for data optimization, and (2) Supporting the deployment of different machine learning frameworks to FPGA requires significant manual effort, consuming a large amount of time. In this paper, we propose AutoCompiler, a unified compiler for mapping NNs to different FPGAs, along with overlay techniques to enable fast and efficient implementation. To the best of our knowledge, we are the first work to support both Deep Neural Networks (DNNs) and Transformer-based networks for overlay-based FPGA deployment. AutoCompiler comprises three integrated enablers: (1) Model Translator, built on top of a topology-based NNs representation, which can optimize the topology and data representation of the models from an algorithmic level based on different hardware configurations, e.g., DSP utilization, (2) Instruction Generator, which generates pipeline data streams according to various FPGA resource configurations by manipulating the instruction set at the upper level rapidly, and (3) End-to-end optimization, which moves as much of the computational processes as possible onto the FPGA chip and minimizes the interaction between CPU and FPGA. Extensive experiments on various Xilinx FPGAs show that AutoCompiler outperforms state-of-the-art overlay-based compiler by 1.2× - 1.35× and same-level GPUs by 1.15× - 1.59× for classic DNN models, and ViT inference, respectively. © 2025 Copyright is held by the owner/author(s). Publication rights licensed to ACM.

关键词： Program compilers

来源：评论

学校读者我要写书评

暂无评论

A data representation method using distance correlation

引用

Frontiers of Computer science 2025年第1期19卷 1-14页

作者： Xinyan LIANG Yuhua QIAN Qian GUO Keyin ZHENG Institute of Big Data Science and Industry Shanxi UniversityTaiyuan 030006China Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of Education Shanxi UniversityTaiyuan 030006China School of Computer Science and Technology Taiyuan University of Science and TechnologyTaiyuan 030024China Shanxi Key Laboratory of Big Data Analysis and Parallel Computing Taiyuan University of Science and TechnologyTaiyuan 030024China

Association in-between features has been demonstrated to improve the representation ability of data. However, the original association data reconstruction method may face two issues: the dimension of reconstructed data is undoubtedly higher than that of original data, and adopted association measure method does not well balance effectiveness and efficiency. To address above two issues, this paper proposes a novel association-based representation improvement method, named as AssoRep. AssoRep first obtains the association between features via distance correlation method that has some advantages than Pearson’s correlation coefficient. Then an improved matrix is formed via stacking the association value of any two features. Next, an improved feature representation is obtained by aggregating the original feature with the enhancement matrix. Finally, the improved feature representation is mapped to a low-dimensional space via principal component analysis. The effectiveness of AssoRep is validated on 120 datasets and the fruits further prefect our previous work on the association data reconstruction.

关键词： association representation distance correlation classification

来源：评论

学校读者我要写书评

暂无评论

Food Images Classification Based on Improved Convolutional Neural Network 4

Food Images Classification Based on Improved Convolutional N...

引用

4th International Conference on Electronic Communication and Artificial Intelligence, ICECAI 2023

作者： He, Yangfan Yin, Siyuan Xiamen University Malaysia School of Computing and Data Science Selangor Malaysia

ISBN: (纸本)9798350314496

Food image classification plays a crucial role in computer vision research and has immense theoretical significance, along with numerous potential applications in various fields such as smart health, smart retail, and smart home. While several studies have been conducted on this topic, there is a lack of consensus on the optimal approach to achieve high accuracy. In this study, the author propose an improved Convolutional Neural Network (CNN) architecture for food image classification. This study aims to enhance the accuracy of food image classification by examining the influence of different optimizers, loss functions, and activation functions on the performance of four neural networks, including VGG19, MobileNet V2, DenseNet, and a traditional CNN architecture, using the ChineseFoodNet dataset. Besides, the paper will improve the performance of the selected model by utilizing data augmentation techniques and transfer learning. The findings of this study can help researchers and practitioners develop more robust and accurate models, facilitating their application in areas such as smart health and smart homes. It also has practical implications for the development of more accurate food image classification models in computer vision field. © 2023 IEEE.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

Optimization of data Protection in IoT Field of Cloud computing Using Compiler Privacy Calculation Method 8

Optimization of Data Protection in IoT Field of Cloud Comput...

引用

8th IEEE International Conference on Software Engineering and Computer Systems, ICSECS 2023

作者： Huang, Jingjia Huang, Xincheng School of Computing and Data Science Xiamen University Malaysia Sepang Malaysia

ISBN: (纸本)9798350310931

This paper aims to enhance cloud computing security and mitigate privacy risks by investigating optimized compiler models. Through extensive literature review and analysis, we have selected a compiler model based on the Porcupine privacy computing approach and a component called Quill, which translates plain text specifications into equivalent implementations in homomorphic cryptography. Concurrently, we propose additional optimization strategies that improve the efficiency and cost effectiveness of homophobic encryption. Our work signifies a step forward in the realm of secure, private cloud computing © 2023 IEEE.

关键词： Cloud computing

来源：评论

学校读者我要写书评

暂无评论

Big data analytics for privacy through ND-homomorphic encryption

引用

Journal of Control and Decision 2023年第1期10卷 64-71页

作者： S.Saravanan N.Poornima Department of CSE SRM Institute of Science and TechnologyChennaiIndia Department of Data Science and Business Systems School of ComputingSRM Institute of Science and TechnologyKattankulathurIndia

Rapidly rising the quantity of Big data is an opportunity to flout the privacy of people. Whenhigh processing capacity and massive storage are required for Big data, distributed networkshave been used. There are several people involved in these activities, the system may contributeto privacy infringements frameworks have been developed for the preservation of privacy atvarious levels (e.g. information age, information the executives and information preparing) asfor the existing pattern of huge information. We plan to frame this paper as a literature surveyof these classifications, including the Privacy Processes in Big data and the presentation of theAssociate Challenges. Homomorphic encryption is particularised aimed at solitary single actionon the ciphered information. Homomorphic enciphering is restrained to an honest operation onthe encoded data. The reference to encryption project fulfils many accurate trading operationson coded numerical data;therefore, it protects the written in code-sensible information evenmore.

关键词： Security and confidentiality vast statistics data encryption nondeterministic fully homomorphic encryption

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：