Vietnamese Dataset Research Articles

XLMR4MD: New Vietnamese dataset and framework for detecting the consistency of description and permission in Android applications using large language models

Google Play and other application marketplaces have various Android applications and metadata. Among these, description information and privacy policy help explain the application's functionality. They also describe the permission of the application, especially those related to sensitive information. Detecting the inconsistency between the description of the application and privacy information and the permission extracted in the application's source code helps users decide whether to install and use the application. In this research, we propose a new method based on a pre-trained language model to detect inconsistencies between the permission extracted from the description application and privacy policy and the permission extracted from the application's source code (file APK). Related works focus on models of large-scale datasets, especially for resource-rich languages such as English. However, a language with low resources, like Vietnamese, needs more datasets for the task. To solve this problem, we propose the ViDPApp dataset (Description and Privacy Policy of Applications on Vietnamese domains), a high-quality dataset that humans manually annotate with 12,000+ sentences with an inter-annotator agreement (IAA) of over 85%. In addition, we proposed XLMR4MD, a new framework using large language models, outperforming powerful machine models (LSTM, Bi-GRU-LSTM-CNN, WikiBERT, DistilBERT, mBERT, and PhoBERT) and achieving the best with 84.04% F1 score in detecting inconsistencies between Android application permission and description. This framework can be fine-tuned for 100 languages, which benefits low-resource languages like Vietnamese. The dataset is available for research purposes.

Fact-checking Vietnamese Information Using Knowledge Graph, Datalog, and KG-BERT

In the era of digital information, ensuring the accuracy and reliability of information is crucial, making fact-checking a vital process. Currently, English fact-checking has thrived due to various language processing tools and ample datasets. However, the same cannot be said for Vietnamese fact-checking, which faces significant challenges due to the lack of such resources. To address these challenges, we propose a model for checking Vietnamese facts by synthesizing three popular technologies: Knowledge Graph (KG), Datalog, and KG-BERT. The KG serves as the foundation for the fact-checking process, containing a dataset of Vietnamese information. Datalog, a logical programming language, is used with inference rules to complete the knowledge within the Vietnamese KG. KG-BERT, a Deep Learning (DL) model, is then trained on this KG to rapidly and accurately classify information that needs fact-checking. Furthermore, to put Vietnamese complex sentences into the fact-checking model, we present a solution for extracting triples from these sentences. This approach also contributes significantly to the ease of constructing foundational datasets for the Vietnamese KG. To evaluate the model's performance, we create a Vietnamese dataset comprising 130,190 samples to populate the KG. Using Datalog, we enrich this graph with additional knowledge. The KG is then utilized to train the KG-BERT model, achieving an impressive accuracy of 95%. Our proposed solution shows great promise for fact-checking Vietnamese information and has the potential to contribute to the development of fact-checking tools and techniques for other languages. Overall, this research makes a significant contribution to the field of data science by providing an accurate solution for fact-checking information in Vietnamese language contexts.

Evaluating the Efficiency of Vietnamese SMS Spam Detection Techniques

Abstract— This paper is aimed at evaluating the efficiency of Vietnamese SMS spam detection methods on different variants of Vietnamese datasets by utilizing both traditional machine learning models and deep learning models. The researchers experimented with five algorithms, which were Support Vector Machine (SVM), Naive Bayes (NB), Random Forests (RF), Convolutional Neural Network (CNN), and Long Short-Term Memory (LSTM), on three different Vietnamese datasets. The findings reveal that the LSTM and CNN, supported by a transformer learning model - PhoBert, were more efficient than the traditional machine learning models. The LSTM model showed the highest accuracy of 97,77% when operating on the full-accent Vietnamese dataset. Similarly, the CNN model and PhoBert model showed the highest accuracy of 95,56% when dealing with non-diacritic Vietnamese dataset.

VinDr-Mammo: A large-scale benchmark dataset for computer-aided diagnosis in full-field digital mammography

Mammography, or breast X-ray imaging, is the most widely used imaging modality to detect cancer and other breast diseases. Recent studies have shown that deep learning-based computer-assisted detection and diagnosis (CADe/x) tools have been developed to support physicians and improve the accuracy of interpreting mammography. A number of large-scale mammography datasets from different populations with various associated annotations and clinical data have been introduced to study the potential of learning-based methods in the field of breast radiology. With the aim to develop more robust and more interpretable support systems in breast imaging, we introduce VinDr-Mammo, a Vietnamese dataset of digital mammography with breast-level assessment and extensive lesion-level annotations, enhancing the diversity of the publicly available mammography data. The dataset consists of 5,000 mammography exams, each of which has four standard views and is double read with disagreement (if any) being resolved by arbitration. The purpose of this dataset is to assess Breast Imaging Reporting and Data System (BI-RADS) and breast density at the individual breast level. In addition, the dataset also provides the category, location, and BI-RADS assessment of non-benign findings. We make VinDr-Mammo publicly available as a new imaging resource to promote advances in developing CADe/x tools for mammography interpretation.

EMPIRICAL STUDY OF FEATURE EXTRACTION APPROACHES FOR IMAGE CAPTIONING IN VIETNAMESE

Image captioning is a challenging task that is still being addressed in the 2020s. The problem has the input as an image, and the output is the generated caption that describes the context of the input image. In this study, I focus on the image captioning problem in Vietnamese. In detail, I present the empirical study of feature extraction approaches using current state-of-the-art object detection methods to represent the images in the model space. Each type of feature is trained with the Transformer-based captioning model. I investigate the effectiveness of different feature types on two Vietnamese datasets: UIT-ViIC and VieCap4H, the two standard benchmark datasets. The experimental results show crucial insight into the feature extraction task for image captioning in Vietnamese.

VLSP 2021-ViMRC Challenge: Vietnamese Machine Reading Comprehension

One of the emerging research trends in natural language understanding is machine reading comprehension (MRC) which is the task to find answers to human questions based on textual data. While many datasets have been developed for MRC research for other languages, there is a lack of such resources for the Vietnamese language. Although many datasets and methodologies have been developed for English and Chinese, many Vietnamese machine reading comprehension limitations need to be solved further. Existing Vietnamese datasets for MRC research concentrate solely on answerable questions. However, in reality, questions can be unanswerable for which the correct answer is not stated in the given textual data. To address the weakness, we provide the research community with a benchmark dataset named UIT-ViQuAD 2.0 for evaluating the MRC task and question answering systems for the Vietnamese language. We use UIT-ViQuAD 2.0 as a benchmark dataset for the shared task on Vietnamese machine reading comprehension (VLSP2021-MRC) at the Eighth Workshop on Vietnamese Language and Speech Processing (VLSP 2021). This task attracted 77 participant teams from 34 universities and other organizations. Each participant was provided with the training data, including 28,457 annotated question-answer pairs, and returned the result on a public test set of more than 3,821 questions and a private test set of 3,712 questions. In this article, we present details of the organization of the shared task, an overview of the methods employed by shared-task participants, and the results. The highest performances in this competition are 77.24% (in EM) and 67.43% (in F1-score) on the private test set. The Vietnamese MRC systems proposed by the top 3 teams use XLM-RoBERTa, a powerful pre-trained language model based on the transformer architecture that has achieved state-of-the-art results on many natural language processing tasks. We believe that releasing the UIT-ViQuAD 2.0 dataset motivates more researchers to improve Vietnamese machine reading comprehension.

Building a Vietnamese Dataset for Natural Language Inference Models.

Natural language inference models are essential resources for many natural language understanding applications. These models are possibly built by training or fine-tuning using deep neural network architectures for state-of-the-art results. That means high-quality annotated datasets are essential for building state-of-the-art models. Therefore, we propose a method to build a Vietnamese dataset for training Vietnamese inference models which work on native Vietnamese texts. Our approach aims at two issues: removing cue marks and ensuring the writing style of Vietnamese texts. If a dataset contains cue marks, the trained models will identify the relationship between a premise and a hypothesis without semantic computation. For evaluation, we fine-tuned a BERT model, viNLI, on our dataset and compared it to a BERT model, viXNLI, which was fine-tuned on XNLI dataset. The viNLI model has an accuracy of 94.79%, while the viXNLI model has an accuracy of 64.04% when testing on our Vietnamese test set. In addition, we also conducted an answer selection experiment with these two models in which the P@1 of viNLI and of viXNLI are 0.4949 and 0.4044, respectively. That means our method can be used to build a high-quality Vietnamese natural language inference dataset.

The Role of Trait Competitiveness and Entrepreneurial Alertness in the Cognitive Process of Entrepreneurship Among Students: A Cross-Cultural Comparative Study Between Vietnam and Poland

Our study aims to explore the influences of trait competitiveness and entrepreneurial alertness on the cognitive process of entrepreneurship in the cross-cultural context of Vietnam and Poland, two emerging nations with different levels of economic and social development. To achieve this research goal, two student questionnaire surveys were carried out at universities and institutes in Vietnam and Poland. Structural equation modelling (SEM) with a bootstrapping approach was utilised to test the proposed hypotheses and conceptual model. Eight hypotheses were statistically supported by the Vietnamese dataset, confirming the significant and positive effects of both trait competitiveness and entrepreneurial alertness on the cognition process of entrepreneurship. However, for the Polish data, trait competitiveness was not found to be associated with an entrepreneurial attitude, perceived behaviour control, or entrepreneurial intention, while entrepreneurial alertness was positively related to perceived behavioural control. Our study has significantly contributed to the entrepreneurship literature by increasing the knowledge about the central role of trait competitiveness and entrepreneurial alertness on the cognitive process of business ventures in two emerging countries, where to the best of our knowledge, few studies related to our topic have been researched. Moreover, practical contributions are also offered for educational institutions and practitioners to stimulate university students’ business venturing activities.

ViReader: A Wikipedia-based Vietnamese reading comprehension system using transfer learning

Machine Reading Comprehension has attracted significant interest in research on natural language understanding, and large-scale datasets and neural network-based methods have been developed for this task. However, most developments of resources and methods in machine reading comprehension have been investigated using two resource-rich languages, English and Chinese. This article proposes a system called ViReader for open-domain machine reading comprehension in Vietnamese by using Wikipedia as the textual knowledge source, where the answer to any particular question is a textual span derived directly from texts on Vietnamese Wikipedia. Our system combines a sentence retriever component, based on techniques of information retrieval to extract the relevant sentences, with a transfer learning-based answer extractor trained to predict answers based on Wikipedia texts. Experiments on multiple datasets for machine reading comprehension in Vietnamese and other languages demonstrate that (1) our ViReader system is highly competitive with prevalent machine learning-based systems, and (2) multi-task learning by using a combination consisting of the sentence retriever and answer extractor is an end-to-end reading comprehension system. The sentence retriever component of our proposed system retrieves the sentences that are most likely to provide the answer response to the given question. The transfer learning-based answer extractor then reads the document from which the sentences have been retrieved, predicts the answer, and returns it to the user. The ViReader system achieves new state-of-the-art performances, with values of 70.83% EM (exact match) and 89.54% F1, outperforming the BERT-based system by 11.55% and 9.54% , respectively. It also obtains state-of-the-art performance on UIT-ViNewsQA (another Vietnamese dataset consisting of online health-domain news) and BiPaR (a bilingual dataset on English and Chinese novel texts). Compared with the BERT-based system, our system achieves significant improvements (in terms of F1) with 7.65% for English and 6.13% for Chinese on the BiPaR dataset. Furthermore, we build a ViReader application programming interface that programmers can employ in Artificial Intelligence applications.

Source Separation using Sparse NMF and Graph Regularization on Vietnamese Dataset

Source separation is popular problem in which English datasets is used by default. Besides, source separation or speech enhancement is an important pre-processing step for following processes e.g. automatic speech recognition, automatic answering machine or hearing ads…However, experiments of source separation on Vietnamese dataset is quite modest as well as lack of Vietnamese standard datasets for source separation. To deal these issues, we build a Vietnamese dataset for source separation by collecting utterances of broadcasters from VTV’s official website. Moreover, a novel method was proposed by using sparse non-negative matrix factorization and graph regularization. Experiments showed that the proposed method is outperformed baseline.

ViMs: a high-quality Vietnamese dataset for abstractive multi-document summarization

Automatic text summarization is important in this era due to the exponential growth of documents available on the Internet. In the Vietnamese language, VietnameseMDS is the only publicly available dataset for this task. Although the dataset has 199 clusters, there are only three documents in each cluster, which is small compared to typical datasets in English. This motivates us to construct ViMs—a big and high-quality Vietnamese dataset for abstractive multi-document summarization. To that end, we recruited 29 annotators and enhanced MDSWriter—an open-source annotation tool, to support the annotators in creating gold standard summaries. As a result, ViMs has 600 summaries corresponding to 300 clusters of 1,945 documents. We have verified the reliability of our dataset by using a variety of metrics including conventional Cohen’s $$\kappa $$ , relaxed Cohen’s $$\kappa $$ —a new metric that we propose to make it more suitable for abstractive summarization, and ROUGE scores. A relaxed $$\kappa $$ score of 0.55 indicate that ViMs could attain moderate agreement between annotators. Meanwhile, ROUGE scores are 0.729 of ROUGE-1, 0.507 of ROUGE-2 and 0.524 of ROUGE-SU4. We have further evaluated ViMs by using three different summarization systems: TextRank, CFVi and MUSEEC. Their performances are 0.628, 0.711 and 0.732 of ROUGE-1, respectively. These results show that the ViMs dataset is suitable for both training and evaluating multi-document summarization systems. We have made the dataset and evaluation results of this work publicly available for research community. It is noted that unlike previous work that only published the final summarization dataset, we also publish intermediate annotation results, which can be used in other NLP problems such as sentence classification.

Genetic classification of Vietnamese cacao cultivars assessed by SNP and SSR markers

Vietnam has the potential to become a larger cocoa producer as the country experiences climatic conditions favorable to the growth of the crop. However, to enable a high cacao quality, genetic methods are required to identify and conserve promising cacao cultivars. Previous research as discussed by Everaert et al. (2017), using 14 microsatellite (SSR) markers, showed a genetic characterization of Vietnamese cacao, but additional research is required to further clarify the genetic background of the different Vietnamese cacao cultivars. Therefore, in this study, the Vietnamese dataset was enlarged, and a classification model was developed. Firstly, we compared the classification strength of the 14 previously applied SSR markers and 42 additional single nucleotide polymorphism (SNP) markers, either separately or combined, using 53 international reference cacao cultivars. Secondly, the most classifying marker set was used to examine the genetic associations of the Vietnamese cultivars to the reference cultivars, using Bayesian clustering and principal coordinate analysis (PCoA). The classification method based on the combined marker set performed best and could subdivide the Vietnamese cultivars in three clusters. The first cluster was strongly related to Nanay, the second to Criollo/Trinitario, and the third predominantly to Amelonado and Criollo/Trinitario. For all marker sets, an apparent high genetic diversity could be observed for the total Vietnam collection. The obtained knowledge provides a strong genetic basis for the Vietnamese cacao industry. In addition, the proposed marker set and the analysis workflow could be applied worldwide for future breeding programs and further production of origin chocolate.

Applying topic model combined with Kohonen networks to discover and visualize communities on social networks

Users are members of communities on social networks. Users’ interested topics keep changing, resulting in the change of their communities’ interested topics as well. Level, period of time, and interested topics represent features of a community which (i) change upon preferences of each user on social networks for making friends or being interested in topics (based on message content); (ii) are formed or change from online groups of friends or the suggestions to make friends. Hence, the link of users in communities can be viewed as a network of users by their features in social network communities. In this paper, the author studies and proposes a new model for discovering communities using Temporal-Author-Recipient-Topic (TART) model combined with Kohonen neural networks to discover communities of users with the same interested topics over different periods of time. The research goal is achieved through testing models on two Vietnamese datasets (collected from social networks at universities and online newspapers).

Association of Leukotriene A4 Hydrolase with Tuberculosis Susceptibility Using Genomic Data in Portugal.

Leukotriene A4 hydrolase (LTA4H) is a key enzyme in the eicosanoid pathway. lta4h locus polymorphisms have previously been linked to tuberculosis (TB) susceptibility and disease outcome in a Vietnamese dataset, but further studies suggested that those results were poorly reproducible. We, therefore, compared the full set of variants (113 SNPs) within the gene in a Portuguese dataset of 112 TB patients and 120 controls, using both the frequency of SNPs and haplotypes, in order to assess their association with TB susceptibility. Although we obtained no significant differences between the TB patients and the control group, linkage analysis showed that an extensively typed polymorphism, rs17525495, was associated with 21 other SNPs, all displaying evidence of association to lower LTA4H expression. While the derived alleles of these SNPs showed a moderately higher frequency in the TB group, differences were not significant. In contrast to Asian populations, where these SNPs are much more frequent, the low frequencies of candidate SNPs in Europeans render them less pertinent in a public health context. Consequently, the typing of specific polymorphisms as a strategy to establish preventive measures and differential TB drug treatments is important but needs to take into consideration that haplotypic background and structure can be substantially different in distinct geographic regions.

A CURE Algorithm for Vietnamese Sentiment Classification in a Parallel Environment

Solutions to process big data are imperative and beneficial for numerous fields of research and commercial applications. Thus, a new model has been proposed in this paper to be used for big data set sentiment classification in the Cloudera parallel network environment. Clustering Using Representatives (CURE), combined with Hadoop MAP (M) / REDUCE (R) in Cloudera – a parallel network system, was used for 20,000 documents in a Vietnamese testing data set. The testing data set included 10,000 positive Vietnamese documents and 10,000 negative ones. After testing our new model on the data set, a 62.92% accuracy rate of sentiment classification was achieved. Although our data set is small, this proposed model is able to process millions of Vietnamese documents, in addition to data in other languages, to shorten the execution time in the distributed environment

Investigation of knowledge transfer approaches to improve the acoustic modeling of Vietnamese ASR system

It is well known that automatic speech recognition &#x0028 ASR &#x0029 is a resource consuming task. It takes sufficient amount of data to train a state-of-the-art deep neural network acoustic model. As for some low-resource languages where scripted speech is difficult to obtain, data sparsity is the main problem that limits the performance of speech recognition system. In this paper, several knowledge transfer methods are investigated to overcome the data sparsity problem with the help of high-resource languages. The first one is a pretraining and fine-tuning &#x0028 PT&#x002F FT&#x0029 method, in which the parameters of hidden layers are initialized with a well trained neural network. Secondly, the progressive neural networks &#x0028 Prognets &#x0029 are investigated. With the help of lateral connections in the network architecture, Prognets are immune to forgetting effect and superior in knowledge transferring. Finally, bottleneck features &#x0028 BNF &#x0029 are extracted using cross-lingual deep neural networks and serves as an enhanced feature to improve the performance of ASR system. Experiments are conducted in a low-resource Vietnamese dataset. The results show that all three methods yield significant gains over the baseline system, and the Prognets acoustic model performs the best. Further improvements can be obtained by combining the Prognets model and bottleneck features.

Minimum Wages and Firm Productivity: Evidence from Vietnamese Manufacturing Firms

ABSTRACTThis paper empirically examines the minimum-wage impact on firm productivity. Using a detailed Vietnamese firm-level dataset from 2010 through 2015, the regression results suggest that firms raise their labor productivity, total factor productivity, capital intensity and revenue in response to increased minimum wage standards. Firms that pay their workers below the minimum wage react more positively in raising their labor productivity than high-wage firms. Minimum wages has had a more pronounced impact on firms’ labor productivity, total factor productivity and capital intensity since the uniform wage rate was introduced for both domestic private and foreign-invested enterprises in 2012.

Sex Differences and Psychological Factors Associated with General Health Examinations Participation: Results from a Vietnamese Cross-Section Dataset

This study focuses on the association of sex differences and psychological factors with periodic general health examination (GHE) behaviors. We conducted a survey in Hanoi and the surrounding areas, collecting 2068 valid observations; the cross-section dataset was then analyzed using the baseline category logit model. The study shows that most people are afraid of discovering diseases through general health examinations (76.64%), and the fear of illness detection appears to be stronger for females than for males (β1(male) = −0.409, p < 0.001). People whose friends/relatives have experienced prolonged treatment tend to show more hesitation in participating in physical check-ups (β2 = 0.221, p < 0.05). On the ideal frequency of GHEs, 90% of the participants agree on once or twice a year. The probability of considering a certain period of time as an appropriate frequency for GHEs changes in accordance with the last doctor visit (low probability of a health examination every 18 months) and one’s fear of potential health problems post-checkup (no fear raises probability of viewing a health examination every 6 months by 9–13 percentage points). The results add to the literature on periodic GHE in particular and on preventive health behaviors in general.

Assessing the Perceptions of the Health Care Workforce Toward the Patient Safety Culture in Vietnamese Hospitals

The main objective of this study was to use the translated version of Hospital Survey on Patient Safety Culture, the English-Vietnamese Hospital Survey on Patient Safety Culture (E-V HSOPSC), to assess the patient safety culture in Vietnamese hospitals and examine the extent to which safety attitudes vary between staff, hospitals, and health care systems. In addition, this study aimed to evaluate the psychometric properties of the instrument in Vietnamese dataset. We evaluated whether patient safety culture contributes to establish a positive patient safety culture—the cornerstone of a quality health care. In May 2015, the E-V HSOPSC was conducted with 1500 staff from 10 hospitals in Vietnam. The respondents were asked to return the completed surveys after a 3-month period. Before assessing the perceptions of health care workforce toward organizational safety culture, a confirmatory factor analysis, construct validity, and reliability were performed using SPSS and Amos 23.0. A total of 1116 questionnaires were eligible for data analysis. The outcomes from factor analysis verified the fitness and validity of the instrument. The positive response rate across 12 safety culture dimensions in the questionnaire ranged from 30% (Hospital Handoffs and Transitions) to 77% (Teamwork within Hospital Units). Overall, the mean positive score was 58.9%, which was slightly lower than of the United States. The safety was graded as “Very Good” by 52.6% of respondents in Vietnam. The E-V HSOPSC was appropriate to assess the patient safety culture in Vietnam, because the instrument provided adequate evidence of validity and reliability and patient safety culture strengths and deficiencies.

Associations of Psychosocial Factors with Maternal Confidence Among Japanese and Vietnamese Mothers

We conducted this cross-sectional study among 392 Japanese and 294 Vietnamese mothers who attended routine child health visits in a Japanese city and at a tertiary hospital in Vietnam, in order to investigate the prevalence and associated sociodemographic, parenting, and psychological characteristics of low maternal confidence in child rearing among them. All data were collected from medical files in Japan, and from medical files and self-administered questionnaires in Vietnam. The proportion of mothers without secure feeling of confidence in the present study was 22% in Japan and 66% in Vietnam. Significant factors associated with a lack of confidence were first-time motherhood and unintended pregnancy in the Japanese dataset and younger age in the Vietnamese dataset. In both groups, a higher proportion of mothers who lacked confidence reported negative parenting outcomes than did confident mothers. Among the three psychological measurements (mood, self-efficacy and depression), higher self-efficacy was associated independently with a significantly reduced risk of not having confidence in child rearing. These results suggest the importance of developing parenting support programs to help Japanese and Vietnamese mothers, particularly those who are young, first-time mothers or who became pregnant unexpectedly, improve their self-efficacy.