检索结果-内蒙古大学图书馆

2024 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2024

作者： Wei, Xiaolin Yang, Ruilong Wu, Xing Wang, Chengliang Wang, Haidong Wang, Hongqian Tang, Tao College of Computer Science Chongqing University Chongqing China Southwest Hospital of Army Medical University Department of Thoracic Surgery Chongqing China Southwest Hospital of Army Medical University Medical Big Data Center Chongqing China

ISBN: (纸本)9781665410205

In focusing tasks on moving targets, traditional methods that rely on maximizing contrast struggle to capture moving objects due to insufficient focusing speed. Deep learning-based methods have attempted to directly predict the optimal focal length for the target;however, due to low prediction accuracy, they often lead to out-of-focus situations when capturing moving objects. In recent years, some approaches have utilized reinforcement learning to automatically explore focal length adjustment patterns, thus achieving better results than traditional methods. However, these approaches have not considered the motion characteristics of the targets, leading to a need for further improvement in focusing performance. To overcome these limitations, we introduce a motion-based feature and deep reinforcement learning-driven autofocus algorithm named MF-DRLAF (Motion Features based Deep Reinforcement Learning Autofocus Model) for moving targets. This novel method tracks the object, predicts its motion state through feature extraction, and uses deep reinforcement learning to dynamically adjust the focus. We utilize a self-attention mechanism to adaptively learn various motion patterns and employ a feature pool structure to enhance processing efficiency. Experiments and real-world testing on a Google Pixel3 demonstrate that our approach significantly enhances autofocus performance on moving objects, highlighting its potential for broader imaging applications. This approach offers a promising direction for future development in autofocus technology. © 2024 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Contrastive Learning based Item Representation with Asymmetric Augmentation for Sequential Recommendation 23

Contrastive Learning based Item Representation with Asymmetr...

引用

2nd International Conference on Algorithms, data Mining, and Information Technology, ADMIT 2023

作者： Zeng, Kexiang Shi, Xiaoyu Xie, Hong Shang, Mingsheng Department of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing China Chongqing Key Laboratory of Big Data and IntelligentComputing Chongqing Institute of Green and Intelligent Technology Chongqing China

ISBN: (纸本)9798400707629

Contrastive learning has been widely applied in sequential recommendation to improve the recommendation performance. Existing contrastive learning methods focus on adjusting the views number of positive and negative samples to enhance the item representation via data-level augmentation(e.g., MMInfoRec). However, they generally ignore the sampled items still follow a long-tail distribution, i.e., a few popular items have a high-frequency occurrence in contrastive learning, while the majority of unpopular items have a low-frequency in contrastive learning. As a result, this imbalance in sample extraction leads to insufficient contrastive learning between popular and unpopular items, resulting in sub-optimal item representation. In this paper, we propose CA4Rec, a Contrastive learning based item representation with Asymmetric augmentation for Sequential Recommendation, that adopts an asymmetric augmented multi-instance contrastive learning strategy to enhance the item representation. Specifically, we first use the popularity-aware method to divide the entire item set into popular and unpopular items, then we generate k augmented views for popular items, and k + n augmented views for unpopular items. Finally, we use our proposed asymmetric multi-instance noise contrastive estimation (AMINCE) to perform contrastive learning calculation. Make sure there is sufficient contrast between popular and unpopular items to mitigate the issue. we implement the improvements on the state-of-the-art model and conduct extensive experiments on three benchmark datasets. The experimental results demonstrate that CA4Rec outperforms the state-of-the-art baselines. © 2023 Owner/Author.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

AARIN: Affordable, accurate, reliable and innovative mechanism to protect a medical cyber-physical system using blockchain technology

引用

International Journal of Intelligent Networks 2021年 2卷 175-183页

作者： Tyagi, Amit Kumar Aswathy, S.U. Aghila, G. Sreenath, N. Centre for Advanced Data Science Vellore Institute of Technology Tamilnadu Chennai600127 India School of Computer Science and Engineering Vellore Institute of Technology Tamilnadu Chennai600127 India Department of Computer Science and Engineering Jyothi Engineering College Kerala Thrissur679531 India Department of Computer Science and Engineering National Institute of Technology Puducherry Karaikal India Department of Computer Science and Engineering Pondicherry Engineering College Pudcuherry India

The enhancement (in the past decade) in electronics devices/technology, a rise in most of the accidents concerning security and surveillance intruding the private lives of the users question the existing systems being used to combat this challenge wherein, the third-parties gather and handle large amounts of individual details. As part of the constant evolution of the cyber physical system architecture, one of the goals of our system is to reduce the latency time for enrollment of new information. Generally, the efficiency and benefit of a Cyber Physical Systems (CPS) depends heavily on interconnection of individual devices or nodes. Exchange of data and information relevant to an overall task or functionality is the key to many applications such as smart grids, smart cities, and many others. Trustworthiness of data is needed to make such systems (especially in MCPS systems) successful. To be able to fulfill policies to guarantee the safety of all entities within a Medical Cyber Physical Systems (MCPS) and to provide security measures to enforce these cryptographic solutions have to be embedded. This paper describes a decentralized e-healthcare application framework for personal data management that ensures that users own and access their data. This work proposes a novel mechanism to secure Medical Cyber Physical Systems (MCPS), i.e., as an automated access-control manager (including building trust in a third party). This work also integrates some features in security building blocks in ultra-small devices to provide essential properties to secure embedded systems. © 2021 The Author(s)

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

Towards a Model of Puzznic

arXiv

引用

arXiv 2023年

作者： Espasa, Joan Gent, Ian P. Miguel, Ian Nightingale, Peter Salamon, András Z. Villaret, Mateu School of Computer Science University of St Andrews United Kingdom Department of Computer Science University of York United Kingdom Department of Computer Science Applied Mathematics and Statistics University of Girona Spain

We report on progress in modelling and solving Puzznic, a video game requiring the player to plan sequences of moves to clear a grid by matching blocks. We focus here on levels with no moving blocks. We compare a planning approach and three constraint programming approaches on a small set of benchmark instances. The planning approach is at present superior to the constraint programming approaches, but we outline proposals for improving the constraint models. © 2023, CC BY.

关键词： Constraint theory

来源：评论

学校读者我要写书评

暂无评论

Improve Deep Learning Autofocus with Depth Information Supervision and Current Focal Distance Cues

Improve Deep Learning Autofocus with Depth Information Super...

引用

2024 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2024

ISBN: (纸本)9781665410205

Traditional autofocus methods search for the optimal focal distance (FD) by evaluating image quality from focal stacks, resulting in time-consuming focusing processes. Recently, deep learning has being adopted for single-shot autofocus methods, which can predict the optimal FD directly from a single input image. However, these methods often suffer from low prediction accuracy due to the lack of global features and structured global supervisory information, as they rely solely on the image's region of interest (ROI) as input and a single value for supervision. We propose a deep learning network named MPFS (Multi-Head Network with Per-Pixel Focal Distance Supervision), which takes a full-frame photograph as input and uses the optimal focal distance per pixel for supervision, this method effectively addresses the issues of missing global features and insufficient supervisory information by leveraging these enhancements. Additionally, the network integrates current camera focal distance information to mitigate the scale ambiguity caused by the lack of absolute scale information. To validate the effectiveness of the proposed method, we designed an experiment using a dataset annotated with optimal FD per pixel. Experimental results on this dataset indicate that our approach achieves a 0.22 decrease in the Mean Absolute Error (Mae) metric compared to the state-of-the-art models, with improvements of 0.02 and 0.004 in d1 and d2 metrics. © 2024 IEEE.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

Don't blame dataset shift! shortcut learning due to gradients and cross entropy 23

Don't blame dataset shift! shortcut learning due to gradient...

引用

Proceedings of the 37th International Conference on Neural Information Processing Systems

作者： Aahlad Puli Lily Zhang Yoav Wald Rajesh Ranganath Department of Computer Science New York University Center for Data Science New York University Department of Computer Science New York University and Center for Data Science New York University and Department of Population Health Langone Health New York University

Common explanations for shortcut learning assume that the shortcut improves prediction under the training distribution but not in the test distribution. Thus, models trained via the typical gradient-based optimization of cross-entropy, which we call default-ERM, utilize the shortcut. However, even when the stable feature determines the label in the training distribution and the shortcut does not provide any additional information, like in perception tasks, default-ERM still exhibits shortcut learning. Why are such solutions preferred when the loss for default-ERM can be driven to zero using the stable feature alone? By studying a linear perception task, we show that default-ERM's preference for maximizing the margin leads to models that depend more on the shortcut than the stable feature, even without overparameterization. This insight suggests that default-ERM's implicit inductive bias towards max-margin is unsuitable for perception tasks. Instead, we develop an inductive bias toward uniform margins and show that this bias guarantees dependence only on the perfect stable feature in the linear perception task. We develop loss functions that encourage uniform-margin solutions, called margin control (MARG-CTRL). MARG-CTRL mitigates shortcut learning on a variety of vision and language tasks, showing that better inductive biases can remove the need for expensive two-stage shortcut-mitigating methods in perception tasks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enhancing Sensor-based Human Activity Recognition Using Hybrid Deep Learning and data Augmentation 5

Enhancing Sensor-based Human Activity Recognition Using Hybr...

引用

5th Research, Invention, and Innovation Congress: Innovative Electricals and Electronics, RI2C 2024

作者： Mekruksavanich, Sakorn Phaphan, Wikanda Jitpattanakul, Anuchit School of Information and Communication Technology University of Phayao Department of Computer Engineering Phayao Thailand King Mongkut's University of Technology North Bangkok Research Group in Statistical Learning and Inference Faculty of Applied Science Department of Applied Statistics Bangkok Thailand King Mongkut's University of Technology North Bangkok Intelligent and Nonlinear Dynamic Innovations Research Center Faculty of Applied Science Department of Mathematics Bangkok Thailand

ISBN: (纸本)9798331505998

Human Activity Recognition (HAR) is essential in various applications, including wellness tracking, automated residences, and fitness monitoring. In the past few decades, sensor-based HAR has become increasingly popular due to advancements in technologies for sensors. Nevertheless, HAR networks' effectiveness dramatically depends on the caliber and volume of the training data, which is frequently restricted and unevenly distributed. This research introduces a novel deep learning method called Multihead-CNN-BiGRU, which integrates one-dimensional convolutional neural networks with bidirectional gated recurrent units (BiGRU) to improve the accuracy of sensor-based HAR. To tackle the problem of insufficient a nd uneven training data, we utilize the synthetic minority over-sampling technique (SMOTE) to augment the data. The suggested model is assessed using the publicly accessible WISDM dataset, which comprises sensor data from diverse human actions. The ID-CNN is employed for extracting localized characteristics from the sensor data, while the BiGRU gathers temporal dependencies and contextual information. The hybrid architecture allows the model to acquire spatial and temporal patterns efficiently. In addition, the SMOTE technique is utilized to create artificial samples of the underrepresented classes, thus equalizing the distribution of classes and enhancing the model's capacity to generalize. The experimental findings show that our hybrid strategy provides exceptional outcomes when paired with SMOTE data augmentation. It obtains the most excellent accuracy of 99.51 % and the highest F1-score of 99.49% compared to the most advanced approaches. The suggested framework provides a reliable and precise solution for sensor-based HAR, which opens up opportunities for improved applications in other fields. © 2024 IEEE.

关键词： Spatio-temporal data

来源：评论

学校读者我要写书评

暂无评论

Addressing the Challenge of Missing Medical data in Healthcare Analytics: A Focus on Machine Learning Predictions for ICU Length of Stay

Addressing the Challenge of Missing Medical Data in Healthca...

引用

2024 Intermountain Engineering, Technology and Computing, IETC 2024

作者： Karankot, Mahmad Isaq Marceau, Max Glenn, Ethan M. Fowers, Rylan P. Hedges, David M. Sheehey, Bonnie Whitaker, Bradley M. Electrical and Computer Engineering Montana State University BozemanMT United States SelectHealth Analytics/Data Science Department MurrayUT United States History and Philosophy Montana State University BozemanMT United States

ISBN: (纸本)9798350373073

This research investigates the impact of missing data on the performance of machine learning algorithms, with a particular focus on the MIMIC-IV dataset. This project aims to investigate the extent to which missing data negatively impacts the training of machine learning algorithms, and whether demographic groups with a higher proportion of missing data (i.e.,ethnicity) have lower predictive accuracy. Using advanced machine learning and data analysis techniques, our results highlight important considerations related to missing data in medical datasets and provide useful insights for improving predictive modeling and decision support systems in clinical practice offers. © 2024 IEEE.

关键词： Singular value decomposition

来源：评论

学校读者我要写书评

暂无评论

A Big data Optimization Comparison using Apache Spark and Apache Airflow

A Big Data Optimization Comparison using Apache Spark and Ap...

引用

IEEE Annual Computing and Communication Workshop and Conference (CCWC)

作者： Taylor Hartman Samir Poudel Jiblal Upadhya Md Nahid Hasan Kritagya Upadhyay Khem Poudel Department of Computer Science Middle Tennessee State University Murfreesboro TN Computational and Data Science Middle Tennessee State University Murfreesboro TN

ISBN: (数字)9798331507695

ISBN: (纸本)9798331507701

While comparisons between Apache Hadoop and Apache Spark are well-documented, there has been limited research comparing Apache Spark with Apache Airflow, especially in terms of speed and memory usage. With Apache Airflow's recent introduction of dynamic task mapping, which performs similar functions to Apache Spark's map operation, a detailed comparison between the two tools has become increasingly relevant. A comparison in these areas would provide valuable insights for the Big data science community, helping determine which methods are better suited for tasks requiring high speed and efficient memory usage. This study focuses on comparing the Apache Spark Map function and Apache Airflow Dynamic Task Mapping function on two key metrics: memory utilization and computation speed. Specifically, we evaluate their performance in sorting formatted electrocardiogram sensory data. We hypoth-esize that Apache Spark will demonstrate faster processing times due to its advanced in-memory processing and sorting algorithms. However, this speed advantage is expected to come with higher memory usage compared to Apache Airflow. Our findings provide actionable insights into the strengths and limitations of these tools, guiding data scientists and engineers in choosing the most suitable framework for specific big data processing tasks. These results are particularly relevant for large-scale data sorting and transformation operations, contributing to informed decision-making in the Big data science community.

关键词： Measurement Hands Machine learning algorithms Heuristic algorithms Pipelines Cluster computing Big data In-memory computing Optimization Sorting

来源：评论

学校读者我要写书评

暂无评论

Time and performance comparison on suicide detection using various feature engineering and machine learning models

Time and performance comparison on suicide detection using v...

引用

International Conference on Knowledge and Smart Technology (KST)

作者： Kittisak Thongsi Nannaphas Booncherd Pokpong Songmuang Department of Data Science and Technology Thammasat Univesity Pathumthani Thailand Department of Computer Science Thammasat Univesity Pathumthani Thailand

Today more people use social media to express their opinion and their emotions. There are many types of text in social media including text that convey a tendency to be depressed or suicidal. We use sentiment analysis to detect suicidal texts, because if detected, it could save many lives and many families. In this research, we have an objective to explore a method that is both high performance and less time-using. We design experiments that have 30 combinations between five machine learning models with six feature engineering methods. All experiments use accuracy and total time for model generation as metrics. We use deep neural networks with glove embedding as a comparator because this combination performed well in this dataset on Kaggle competition. From the experimental results, we find that the suitable combination that generates fast and has good accuracy is Random Forest with TF-IDF with 0.897 and 145 seconds.

关键词： Measurement Knowledge engineering Deep learning Sentiment analysis Social networking (online) Text categorization Neural networks

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：