检索结果-内蒙古大学图书馆

Ensemble probabilistic quantization encoding for information preservation of numerical variables in convolutional neural networks

引用

SCIENTIFIC REPORTS 2025年第1期15卷 1-13页

作者： Nam, Ki Yup Park, Hyun-Woong Lee, Yeongseop Jung, Hun-Young Kang, Taeseen Chungnam Natl Univ Hosp Dept Ophthalmol Daejeon South Korea Chungnam Natl Univ Inst Hlth Sci Dept Ophthalmol Coll Med Daejeon South Korea Chungnam Natl Univ Sejong Hosp Dept Cardiol Sejong South Korea Designovel Seoul South Korea Chungnam Natl Univ Dept Ophthalmol Sejong Hosp 20 Bodeum 7 Ro Sejong South Korea

One-hot encoding is a prevalent method used to convert numeric variables into categorical variables. But one-hot encoding omits crucial quantitative data, which compromises the performance of convolutional neural networks (CNNs). This study introduces the ensemble probabilistic quantization encoding, where each class is treated as a quantum with distinct values and the classes collaborate in an ensemble manner to preserve numerical information. This method uses the cross-entropy loss function, enhancing its robustness to outliers. Moreover, classes collaborate in an ensemble fashion to yield more diverse and enriched outcomes. We compared three encoding techniques-ensemble probabilistic quantization, one-hot encoding, label smoothing, and mean square error-using the same dataset and model structure. Our investigations into the impact of quantitative information loss on CNN performance revealed that omitting this information significantly undermines CNN functionality. Ensemble probabilistic quantization proved less dependent on the number of classes than the other methods, thus maintaining effectiveness even with fewer classes. In conclusion, the efficient transmission of quantitative information from numerical to categorical variables is essential for optimal CNN performance. Ensemble probabilistic quantization effectively conveys diverse quantitative information with fewer classes, outperforming one-hot encoding and label smoothing when class numbers are limited.

关键词： Categorical variables Convolutional neural networks encoding methods Numerical variables

来源：评论

学校读者我要写书评

暂无评论

Performance and Optimization of Coding Techniques for Deep Space Communication Channels 7

Performance and Optimization of Coding Techniques for Deep S...

引用

7th International Conference on Automation Electronics and Electrical Engineering

作者： Guan, Wenqi Nanjing Univ Sch Elect Sci & Engn Nanjing Peoples R China

ISBN: (纸本)9798350377040;9798350377033

In today's world, space exploration and communication advancements are critical to ensuring effective connections for scientific missions, planetary exploration, and future interstellar travel. Deep space communication plays a pivotal role in supporting these endeavors, as it enables the transmission of vital data across vast distances and harsh environments. This paper presents a comprehensive study on the performance and optimization of coding techniques for deep-space communication channels. Based on the characteristics of the deep-space environment, the channels are modeled into three stages: near-Earth link, interstellar link, and near-planet link. The modeling for each channel type is summarized, and simulation results are provided. To evaluate coding techniques, M2M4, MMSE, and DNN are employed to approximate the signal-to-noise ratio (SNR), which serves as a critical parameter in assessing various encoding schemes. Three coding techniques-Hamming, convolutional, and turbo codes-are tested in the context of these models, with a particular focus on the turbo coding method. Detailed exploration is carried out on the SOVA and Log-MAP algorithms, as well as the impact of transmission parameters such as code rate. Overall, the results aim to provide a foundation for reliable signal transmission and error correction in deep-space, addressing challenges inherent to the unique environment of deep-space communication.

关键词： deep-space communication channel modeling encoding methods SNR estimation hard and soft decisions

来源：评论

学校读者我要写书评

暂无评论

Nonlinear Wavefront Shaping and Imaging in Nonlinear Photonic Crystals with a Generalized Quasi-Multivalue-encoding Method

引用

ACS PHOTONICS 2021年第11期8卷 3133-3140页

作者： Lu, Ronger Zhao, Ruizhi Feng, Xia Zhang, Chao Hong, Xuhao Qin, Yiqiang Zhu, Yongyuan Nanjing Univ Natl Lab Solid State Microstruct Nanjing 210093 Peoples R China Nanjing Tech Univ Dept Phys Nanjing 211816 Peoples R China Nanjing Univ Coll Engn & Appl Sci Nanjing 210093 Peoples R China Nanjing Univ Sch Phys Nanjing 210093 Peoples R China

Nonlinear photonic crystals have been employed as computer-generated holograms to manipulate the wavefront of the nonlinear waves. However, the disadvantages of the binary-fabricated structures gradually appear because of their adverse effects on the manipulation quality, especially in the field of nonlinear shaping and imaging. A generalized encoding method has been proposed for constructing nonlinear holographic processes of high quality, which is called a quasi-multivalue-encoding method. Both theoretical and experimental results demonstrate that, compared with the conventional binary encoding method, this generalized encoding method is helpful to reduce the distortion and high-order diffraction effects. The integrated ability to convert the light frequency and to reshape its wavefront with a high resolution in any desired way can be very useful for all-optical shaping, switching, imaging, manipulating, and so on.

关键词： second-harmonic generation nonlinear photonic crystals shaping and imaging encoding methods high resolution

来源：评论

学校读者我要写书评

暂无评论

Exploring potential biosafety implications in DNA information storage

引用

BIOSAFETY AND HEALTH 2025年第2期7卷 132-139页

作者： Li, Shangzhe Shi, Yue Yang, Jing Liu, Haizhou Jia, Lijia Liu, Di Chinese Acad Sci Wuhan Inst Virol Ctr Biosafety Megasci Key Lab Special Pathogens & Biosafety Wuhan 430071 Peoples R China Univ Chinese Acad Sci Beijing 101408 Peoples R China

Deoxyribonucleic acid (DNA) information storage has emerged as a promising solution to address the challenges of traditional silicon-based data storage systems. However, the biosafety implications of artificially synthesized DNA sequences in this technology remain understudied. This research evaluates the biosafety risks associated with five representative DNA storage encoding methods [Church, Goldman, DNA Fountain, Grass, and movable-type (MT) encoding] by analyzing their sequence similarities to natural biological DNA. Through Kraken2 taxonomic classification and Basic Local Alignment Search Tool for nucleotides (BLASTn) alignment analysis, we found that while most artificially designed DNA sequences showed significant differences from known biological sequences, specific encoding methods produced sequences similar to natural genomes. The MT encoding method showed the highest annotation rate (4.59 %) in Kraken2 analysis, while Goldman and Fountain methods demonstrated significant local sequence alignments in BLASTn analysis. Sequence length positively correlated with annotation rates, suggesting longer sequences pose potentially higher biosafety risks. Furthermore, aligned sequences often exhibited characteristics of tandem repeats, particularly in non-coding regions. These findings highlight the importance of incorporating biosafety considerations in DNA storage encoding method development and suggest that randomization strategies may help mitigate potential risks. Our study provides valuable insights into the safe advancement of DNA storage technology and emphasizes the need for comprehensive biosafety evaluation in synthetic biology applications. (c) 2025 Published by Elsevier B.V. on behalf of Chinese Medical Association Publishing House. This is an open access article under the CC BY-NC-ND license (http://***/licenses/by-nc-nd/4.0/).

关键词： DNA information storage Biosafety Biosecurity encoding methods Synthetic DNA

来源：评论

学校读者我要写书评

暂无评论

Combining discrete choice models and neural networks through embeddings: Formulation, interpretability and performance

引用

TRANSPORTATION RESEARCH PART B-METHODOLOGICAL 2023年第1期175卷

作者： Arkoudi, Ioanna Krueger, Rico Azevedo, Carlos Lima Pereira, Francisco C. DTU Tech Univ Denmark Roskilde Denmark

This paper proposes a novel approach that combines theory-and data-driven choice models using Artificial Neural Networks (ANNs). In particular, we use continuous vector represen-tations, called embeddings, for encoding categorical or discrete explanatory variables with a special focus on interpretability and model transparency. Although embedding representations within the logit framework have been conceptualized by Pereira (2019), their dimensions do not have an absolute definitive meaning, hence offering limited behavioral insights in this earlier work. The novelty of our work lies in enforcing interpretability to the embedding vectors by formally associating each of their dimensions to a choice alternative. Thus, our approach brings benefits much beyond a simple parsimonious representation improvement over dummy encoding, as it provides behaviorally meaningful outputs that can be used in travel demand analysis and policy decisions. Additionally, in contrast to previously suggested ANN-based Discrete Choice Models (DCMs) that either sacrifice interpretability for performance or are only partially interpretable, our main model (E-MNL) preserves interpretability of the utility coefficients for all the input variables despite being based on ANN principles. The proposed models are tested on two real world datasets and evaluated against benchmark and baseline models that use dummy-encoding. The results of the experiments indicate that our models deliver state-of-the-art predictive performance, outperforming existing ANN-based models while drastically reducing the number of required network parameters.

关键词： Categorical embeddings encoding methods Interpretable embeddings Transparent neural networks Behavioral modeling Discrete choice models

来源：评论

学校读者我要写书评

暂无评论

Trace encoding in process mining: A survey and benchmarking

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2023年第PartD期126卷

作者： Tavares, Gabriel M. Oyamada, Rafael S. Barbon Junior, Sylvio Ceravolo, Paolo Ludwig Maximilians Univ Munchen Oettingenstr 67 D-80538 Munich Germany Univ Milan Via Giovanni Celoria 20 I-20133 Milan Italy Univ Trieste Piazzale Europa 1 I-34127 Trieste Italy

encoding methods are employed across several process mining tasks, including predictive process monitoring, anomalous case detection, trace clustering, etc. These methods are usually performed as preprocessing steps and are responsible for mapping complex event data information into a numerical feature space. Most papers choose existing encoding methods arbitrarily or employ a strategy based on expert domain knowledge. Moreover, existing methods are employed by using their default parameters without evaluating other options. This practice can lead to several drawbacks, such as suboptimal performance and unfair comparisons with the state-of-the-art. Therefore, this work aims at providing a comprehensive survey and benchmark on event log encoding by comparing 27 methods, from different natures, in terms of expressivity, scalability, correlation, and domain agnosticism. To the best of our knowledge, this is the most comprehensive study so far focusing on trace encoding in process mining. It contributes to maturing awareness about the role of trace encoding in process mining pipelines and sheds light on issues, concerns, and future research directions regarding the use of encoding methods to bridge the gap between machine learning models and process mining.

关键词： encoding methods Process mining Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Exploring potential biosafety implications in DNA information storage

引用

生物安全与健康(英文) 2025年第2期

作者： Li Shangzhe Shi Yue Yang Jing Liu Haizhou Jia Lijia Liu Di Key Laboratory of Special Pathogens and Biosafety Center for Biosafety Mega-Science Wuhan Institute of Virology Chinese Academy of Sciences Wuhan China University of Chinese Academy of Sciences Beijing China

关键词： DNA information storage Biosafety Biosecurity encoding methods Synthetic DNA

来源：评论

学校读者我要写书评

暂无评论

XML Labels Compression using Prefix-encodings 12

XML Labels Compression using Prefix-encodings

引用

12th International Conference on Web Information Systems and Technologies (WEBIST)

作者： Al Zadjali, Hanaa North, Siobhan Univ Sheffield Dept Comp Sci Sheffield S Yorkshire England

ISBN: (纸本)9789897581861

XML is the de-facto standard for data representation and communication over the web, and so there is a lot of interest in querying XML data and most approaches require the data to be labelled to indicate structural relationships between elements. This is simple when the data does not change but complex when it does. In the day-to-day management of XML databases over the web, it is usual that more information is inserted over time than deleted. Frequent insertions can lead to large labels which have a detrimental impact on query performance and can cause overflow problems. Many researchers have shown that prefix encoding usually gives the highest compression ratio in comparison to other encoding schemes. Nonetheless, none of the existing prefix encoding methods has been applied to XML labels. This research investigates compressing XML labels via different prefix-encoding methods in order to reduce the occurrence of any overflow problems and improve query performance. The paper also presents a comparison between the performances of several prefix-encodings in terms of encoding/decoding time and compressed code size.

关键词： XML Labels XML Compression encoding methods Prefix encoding

来源：评论

学校读者我要写书评

暂无评论

Supporting Orthographic Learning at the Beginning Stage of Learning to Read Chinese as a Second Language

引用

The Exceptional Child 2014年第3期61卷 288-305页

Learning to read a second language (L2) is especially challenging when a target L2 requires learning new graphic forms. Learning Chinese, which consists of thousands of characters composed of hundreds of basic writing units, presents such a challenge of orthographic learning for adult English speakers at the beginning stages of learning. In this study, we use anin vivoclassroom design to extend previous research on how to support orthographic learning. First, we test the hypothesis that learning characters is enhanced by a grouped sequence of characters that share sub-character graphic components. Second, we examine the effects of four encoding methods that have been investigated in laboratory studies—handwriting, visual chunking, passive reading, and stroke-reporting. The results demonstrate that the grouped approach facilitated character production compared with the distributed approach and that visual-chunking outperformed the other three encoding methods under the grouped sequence. We propose that learning via visual chunking with characters grouped by the same chunks enhances the Chinese orthographic representations of beginning L2 learners.

关键词： characters-distributed instruction characters-in-group instruction character scoring scheme Chinese as foreign/second-language learning encoding methods handwriting in vivo experiment orthographic learning reading visual chunking

来源：评论

学校读者我要写书评

暂无评论

Optimization of the parameters of an All-optical Chaotic Communication System - art. no. 60211J

Optimization of the parameters of an All-optical Chaotic Com...

引用

Conference on Optical Transmission, Switching and Subsystems III

作者： Chang, DY Zheng, K Ma, LN Chen, T Jian, SS Beijing Jiaotong Univ Inst Lightwave Technol Beijing 100044 Peoples R China

ISBN: (纸本)0819460524

Kanakidis et al presented several kinds of all-optical chaotic communication systems using two encoding techniques and various dispersion compensation maps [1]. It shows that the permitted transmission distances are different for various dispersion compensation maps and various encoding techniques. In order to explore the upper limits of the transmission distance, the parameters of the all-optical chaotic communication system introduced by D. Kanakidis et al. [1] is optimized using genetic algorithm. Some useful results are presented.

关键词： chaotic communication genetic algorithm encoding methods dispersion management transmission

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：