2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT)最新文献

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00146

L. Jiyun, Frimpong Felix, Wang Yifan

{"title":"Patient Similarity Measuring with Graph Embedded Learning and Triplet Network","authors":"L. Jiyun, Frimpong Felix, Wang Yifan","doi":"10.1109/CECIT53797.2021.00146","DOIUrl":"https://doi.org/10.1109/CECIT53797.2021.00146","url":null,"abstract":"Electronic Health Records (EHR) are fast becoming the standard for patient medical record keeping. Patient similarity measuring enabled by EHRs has demonstrated the ability to assist doctors and other medical practitioners in clinical decision-making from prognosis, to diagnosis, to treatment options and treatment. Due to the complexity and heterogeneity of medical data, it remains a challenge to integrate critical latent information to estimate patient similarity. As a result, nuances like the ability of a medical entity to have multiple meanings in different contexts and its effect on whether patients are similar or not, a key player in mitigating misdiagnosis, remain a relevant problem. We seek to take full advantage of the usually heterogeneous data of EHRs, to more carefully measure patient similarity by modeling the medical entities they contain as well as their relationships both with the patient and with other medical entities in the patient EHR scenarios through the use of heterogeneous graph embed-dings and triplet networks. We demonstrate the effectiveness of our algorithm through experiments on breast cancer diagnosis and patient medical record data from a tertiary hospital.","PeriodicalId":214527,"journal":{"name":"2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT)","volume":"198 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115010958","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Recognizing Sentences Concerning Research Methods of Citation from the Full Text of JASIST 《JASIST》全文引文研究方法辨析

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00160

B. Tan, Dongbo Wang, Si Shen

引用次数: 0

READ-CSRL: Relation Entities Augmented Datasets for Conversational Semantic Role Labeling 会话语义角色标注的关系实体增强数据集

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00157

Yinyan Wu, Chenhao Jin, Congduan Li

{"title":"READ-CSRL: Relation Entities Augmented Datasets for Conversational Semantic Role Labeling","authors":"Yinyan Wu, Chenhao Jin, Congduan Li","doi":"10.1109/CECIT53797.2021.00157","DOIUrl":"https://doi.org/10.1109/CECIT53797.2021.00157","url":null,"abstract":"Some intelligent applications such as intelligent customer service, in-depth Q&A, chat bot, etc. need multiple conversation rounds, and there is a strong correlation between conversational sentences. Traditional Semantic Role Labeling (SRL) can only perform well on single-sentence tasks, however, the performances on tasks with multiple conversational sentences are unsatisfactory. Conversational Semantic Role Labeling (CSRL) was proposed for conversational tasks. Then knowledge enhanced CSRL was proposed, commonsense was introduced as external knowledge which is usually represented as a triple of request entity, relation and relation entity. Experiments show that after knowledge enhanced by full triples may bring some noise into the original input. So, in this paper, we propose a novel method to reduce the knowledge noise, and generate more augmented datasets enhanced with only relation entities. Experiments evidently illustrate that the proposed method outperforms the CSRL benchmarks. This provides a promising new way to fuse external knowledge in conversational SRL task.","PeriodicalId":214527,"journal":{"name":"2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT)","volume":"384 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115609017","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An Self-supervised Learning & Self-attention Based Method for Defects Classification on PCB Surface Images 基于自监督学习和自注意的PCB表面图像缺陷分类方法

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00047

Lijie Zhou, Xufeng Ling, Shan Zhu, Zheng Sun, Jie Yang

{"title":"An Self-supervised Learning & Self-attention Based Method for Defects Classification on PCB Surface Images","authors":"Lijie Zhou, Xufeng Ling, Shan Zhu, Zheng Sun, Jie Yang","doi":"10.1109/CECIT53797.2021.00047","DOIUrl":"https://doi.org/10.1109/CECIT53797.2021.00047","url":null,"abstract":"Automatic detection of PCB defects become a difficult work in electronics industry, with the rapid development of Integrated Circuit. A good detection method can effectively improve production efficiency and reduce the costs of quality inspection. The traditional image processing methods have the backwards of complicated system structure, difficult image segmentation, and low detection accuracy, etc. In order to solve the above problems, this paper proposes a method of extracting features by self-supervised learning with supervised learning to complete the classification task. This method uses a deep neural network model based on self-supervised learning technology, uses multi-head self-attention mechanism; automatically recognize single or multiple defect areas in PCB, to extract defects features of PCB, and uses full connection classifier to classify features. The open-source (DeepPCB) is selected as the dataset, including 6 defect categories and non-defect categories. In the training process, data enhancement technology is used to increase the training samples, train the model and make it converge, and realize low error recognition on the verification dataset. The experimental results show that the features extracted by the self-supervised learning method are richer than those extracted by the supervised learning. The composite model formed by combining the self-attention mechanism has high accuracy and good robustness, reaches the classification accuracy of 94.33%, and has high accuracy and good generalization ability.","PeriodicalId":214527,"journal":{"name":"2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115736387","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 4

Improving Discriminative Entity Retriever with Generative Tasks 基于生成任务的判别实体检索算法改进

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00120

Siyu Lv

{"title":"Improving Discriminative Entity Retriever with Generative Tasks","authors":"Siyu Lv","doi":"10.1109/CECIT53797.2021.00120","DOIUrl":"https://doi.org/10.1109/CECIT53797.2021.00120","url":null,"abstract":"Entity retrieval, which aims to retrieve a subset of candidate entities from massive KBs for each query, is essential for many entity-centric tasks, e.g., question answering and entity linking. Most of existing methods address this task using a discriminative model by formulating it as as a matching problem between the query and each candidate entity. Recent progress show that the generative model, which directly generate entity names in an autoregressive way, can obtain competitive results against discriminative models. These two types of methods model the interaction between queries and entities in different fashion, where both of them have distinct advantages. In this work, we attempt to join the discriminative and generative modeling to leverage the benefits of both. Specifically, we develop a framework to jointly train a text generative model and a discriminative retrieval model, we use two-tower structure to encode queries and entities respectively, and use Bert as our encoder. Our model achieves 0.14 improvement in NDCG@10. And we compare the results of our model with the model without generative model, and the experiment proves that the generative model is effective.","PeriodicalId":214527,"journal":{"name":"2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124879962","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Research on the Construction Method of Zhuang Word Embedding Evaluation Set 壮词嵌入评价集的构建方法研究

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00188

Yunyou Huang, Danni He, Suqin Tang, Tinghui Li

{"title":"Research on the Construction Method of Zhuang Word Embedding Evaluation Set","authors":"Yunyou Huang, Danni He, Suqin Tang, Tinghui Li","doi":"10.1109/CECIT53797.2021.00188","DOIUrl":"https://doi.org/10.1109/CECIT53797.2021.00188","url":null,"abstract":"Word embedding is the first process to study the natural language processing. However, most of the researches on word embedding of minority languages, such as Zhuang language, lacks a word embedding evaluation method since the cost of manually constructing the test set of word embedding is high, and there are few minority language applications. In addition, the test set of word embedding of minority languages constructed by the traditional method usually lacks diversity. This paper proposes an improved sampling algorithm SPXY_Word to automatically create the comprehensive minority word embedding test set. Taking Zhuang as an example, we realize the automatic construction of the Zhuang word embedding test set Zhuang-Benchmark900. The experimental results show that the test set has balanced data, good discrimination between different word embedding models, highly similar evaluation ability with the Chinese test set: taking Pearson coefficient as the evaluation index, the similar evaluation ability of the two test sets is 97% in similarity evaluation and is 60% in relevance evaluation. This study provides a new solution for constructing a word embedding evaluation set for minority languages, which is helpful to slow the decline of minority languages and cultures.","PeriodicalId":214527,"journal":{"name":"2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT)","volume":"185 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125167106","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The Decoding Method of LDPC based on Dynamic Maximum Iteration in Physical Downlink Shared Channel data 基于动态最大迭代的LDPC物理下行共享信道数据解码方法

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00213

Liang Yan

引用次数: 0

An PM2.5-driven parametric adaptive ESDD prediction model 一种pm2.5驱动的参数自适应ESDD预测模型

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00046

Yi Ma, Bo Wu, Dada Wang, Fangrong Zhou, Hao Pan, G. Wen, Chao Xu, Zhidong Li

{"title":"An PM2.5-driven parametric adaptive ESDD prediction model","authors":"Yi Ma, Bo Wu, Dada Wang, Fangrong Zhou, Hao Pan, G. Wen, Chao Xu, Zhidong Li","doi":"10.1109/CECIT53797.2021.00046","DOIUrl":"https://doi.org/10.1109/CECIT53797.2021.00046","url":null,"abstract":"ESDD is the main cause of high voltage transmission line ESDD and flashing accidents, and it is of great importance to study the accumulation law of insulator ESDD. The existing ESDD prediction model assumes that the rate of ESDD accumulation is a relatively constant quantity, and the observed data show that this assumption does not hold true in China. The rate of ESDD accumulation is closely related to the air pollution, and our study establishes the relationship between the rate of ESDD accumulation and the particulate matter 2.5 (PM2.5). The dynamic prediction model of dirt accumulation was constructed on this basis, and the idea of data assimilation was introduced to complete the parameter optimization estimation. The experimental results show that the model can not only describe the dynamic change law of the equivalent salt density well, but also make more accurate and stable prediction of the equivalent salt density. Considering the easy accessibility and large wide coverage of PM2.5 data, the proposed model is expected to significantly reduce the cost of regional ESDD assessment.","PeriodicalId":214527,"journal":{"name":"2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT)","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116431522","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Performance Comparison of ECDH and ECDSA ECDH与ECDSA的性能比较

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00149

M. Koppl, D. Siroshtan, M. Orgon, Stefan Pocarovsky, Antonin Bohacik, Karel Kuchar, E. Holasova

引用次数: 1

Civil Aviation Data Monitoring System Based on Encryptable Consortium Blockchain 基于加密联盟区块链的民航数据监控系统

2021 2nd International Conference on Electronics, Communications and Information Technology (CECIT) Pub Date : 2021-12-01 DOI: 10.1109/CECIT53797.2021.00074

Hang Zhu, Yu Su, Yuhe Qiu, Jing Li, Hao Xu, Xuhan Liao

引用次数: 2