Pub Date : 2023-08-07DOI: 10.25073/2588-1086/vnucsce.662
D. Thin, N. Nguyen
In recent years, Aspect-based sentiment analysis (ABSA) has received increasing attention from the scientific community for Vietnamese language. However, most previous studies solved various subtasks in ABSA based on machine learning, deep learning and transformer-based architectures in a classification way. Recently, the release of pre-trained sequence-to-sequence brings a new approach to address the ABSA as a text generation problem for Vietnamese ABSA tasks. In this paper, we formulate the Aspect-category based sentiment analysis task as the conditional text generation task and investigate different unified generative transformer-based models. To represent the labels in a natural sentence, we apply a simple statistical method and observation of the commenting style. We conduct experiments on two benchmark datasets. As a result, our model achieved new state-of-the-art results with the micro F1-score of 75.53% and 86.60% for the two datasets with different levels for the restaurant domain. In addition, our experimental results achieved the best score for the smartphone domain with the macro F1-score of 81.10%.
{"title":"Aspect-Category based Sentiment Analysis with Unified Sequence-To-Sequence Transfer Transformers","authors":"D. Thin, N. Nguyen","doi":"10.25073/2588-1086/vnucsce.662","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.662","url":null,"abstract":"In recent years, Aspect-based sentiment analysis (ABSA) has received increasing attention from the scientific community for Vietnamese language. However, most previous studies solved various subtasks in ABSA based on machine learning, deep learning and transformer-based architectures in a classification way. Recently, the release of pre-trained sequence-to-sequence brings a new approach to address the ABSA as a text generation problem for Vietnamese ABSA tasks. In this paper, we formulate the Aspect-category based sentiment analysis task as the conditional text generation task and investigate different unified generative transformer-based models. To represent the labels in a natural sentence, we apply a simple statistical method and observation of the commenting style. We conduct experiments on two benchmark datasets. As a result, our model achieved new state-of-the-art results with the micro F1-score of 75.53% and 86.60% for the two datasets with different levels for the restaurant domain. In addition, our experimental results achieved the best score for the smartphone domain with the macro F1-score of 81.10%.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121241789","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}
Pub Date : 2023-08-07DOI: 10.25073/2588-1086/vnucsce.596
Hung K. Nguyen, Long Quoc Tran
The computation complexity and huge memory access bandwidth of the convolutional layers in convolutional neural networks (CNNs) require specialized hardware architectures to accelerate CNN’s computations while keeping hardware costs reasonable for area-constrained embedded applications. This paper presents an RTL (Register Transfer Logic) level microarchitecture of hardware- and bandwidth-efficient high-performance 2D convolution unit for CNN in deep learning. The 2D convolution unit is made up of three main components including a dedicated Loader, a Circle Buffer, and a MAC (Multiplier-Accumulator) unit. The 2D convolution unit has a 2-stage pipeline structure that reduces latency, increases processing throughput, and reduces power consumption. The architecture proposed in the paper eliminates the reloading of both the weights as well as the input image data. The 2D convolution unit is configurable to support 2D convolution operations with different sizes of input image matrix and kernel filter. The architecture can reduce memory access time and power as well as execution time thanks to the efficient reuse of the preloaded input data, while simplifying hardware implementation. The 2D convolution unit has been simulated and implemented on Xilinx's FPGA platform to evaluate its superiority. Experimental results show that our design is 1.54× and 13.6× faster in performance than the design in [7] and [8], respectively, at lower hardware cost without using any FPGA’s dedicated hardware blocks. By reusing preloaded data, our design achieves a bandwidth reduction ratio between 66.4% and 90.5%.
{"title":"A Bandwidth-Efficient High-Performance RTL-Microarchitecture of 2D-Convolution for Deep Neural Networks","authors":"Hung K. Nguyen, Long Quoc Tran","doi":"10.25073/2588-1086/vnucsce.596","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.596","url":null,"abstract":"The computation complexity and huge memory access bandwidth of the convolutional layers in convolutional neural networks (CNNs) require specialized hardware architectures to accelerate CNN’s computations while keeping hardware costs reasonable for area-constrained embedded applications. This paper presents an RTL (Register Transfer Logic) level microarchitecture of hardware- and bandwidth-efficient high-performance 2D convolution unit for CNN in deep learning. The 2D convolution unit is made up of three main components including a dedicated Loader, a Circle Buffer, and a MAC (Multiplier-Accumulator) unit. The 2D convolution unit has a 2-stage pipeline structure that reduces latency, increases processing throughput, and reduces power consumption. The architecture proposed in the paper eliminates the reloading of both the weights as well as the input image data. The 2D convolution unit is configurable to support 2D convolution operations with different sizes of input image matrix and kernel filter. The architecture can reduce memory access time and power as well as execution time thanks to the efficient reuse of the preloaded input data, while simplifying hardware implementation. The 2D convolution unit has been simulated and implemented on Xilinx's FPGA platform to evaluate its superiority. Experimental results show that our design is 1.54× and 13.6× faster in performance than the design in [7] and [8], respectively, at lower hardware cost without using any FPGA’s dedicated hardware blocks. By reusing preloaded data, our design achieves a bandwidth reduction ratio between 66.4% and 90.5%.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125919106","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}
Pub Date : 2023-06-16DOI: 10.25073/2588-1086/vnucsce.659
Viet-Cuong Ta, Thu Uyen Do
In this paper, we study the efficiency of Graph Transformer Network for noisy label propagation in the task of classifying video anomaly actions. Given a weak supervised dataset, our methods focus on improving the quality of generated labels and use the labels for training a video classifier with deep network. From a full-length video, the anomaly properties of each segmented video can be decided through their relationship with other video. Therefore, we employ a label propagation mechanism with Graph Transformer Network. Our network combines both the feature-based relationship and temporal-based relationship to project the output features of the anomaly video to a hidden dimension. By learning in the new dimension, the video classifier can improve the quality of noisy, generated labels. Our experiments on three benchmark dataset show that the accuracy of our methods are better and more stable than other tested baselines.
{"title":"Noisy-label propagation for Video Anomaly Detection with Graph Transformer Network","authors":"Viet-Cuong Ta, Thu Uyen Do","doi":"10.25073/2588-1086/vnucsce.659","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.659","url":null,"abstract":"In this paper, we study the efficiency of Graph Transformer Network for noisy label propagation in the task of classifying video anomaly actions. Given a weak supervised dataset, our methods focus on improving the quality of generated labels and use the labels for training a video classifier with deep network. From a full-length video, the anomaly properties of each segmented video can be decided through their relationship with other video. Therefore, we employ a label propagation mechanism with Graph Transformer Network. Our network combines both the feature-based relationship and temporal-based relationship to project the output features of the anomaly video to a hidden dimension. By learning in the new dimension, the video classifier can improve the quality of noisy, generated labels. Our experiments on three benchmark dataset show that the accuracy of our methods are better and more stable than other tested baselines.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129418158","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}
Pub Date : 2023-06-06DOI: 10.25073/2588-1086/vnucsce.803
Duc-Hanh Dang
In software development, to obtain a precise specification of the software system's functional requirements is significant to ensure the software quality as well as to automate the development. Use cases are an effective way to capture functional requirements, however, the use of ambiguous or vague language in the use case can lead to imprecision. It is essential to ensure that use case specifications are clear, concise, and complete to avoid imprecision in requirements. This paper aims to develop a domain specific language called FRSL to precisely specify use cases and to provide a basis for transformations to generate software artifacts from the use case specification. We define a FRSL metamodel to capture the technical domain of use cases for FRSL's abstract syntax, and then provides a textual concrete syntax for this language. We also define a formal operational semantics for FRSL by characterizing the execution of a FRSL specification as sequences of system snapshot transitions. This formal semantics on the one hand allows us to precisely explain the meaning of use cases and their relationships, on the other hand provides a basis for transformations from the use case specification. We implement a tool support for this language and perform an evaluation of its expressiveness in comparison with current use case specification languages. This work brings out (1)~a DSL to specify use cases that is defined based on a formal semantics of use cases; and (2)~a tool support realized as an Eclipse plugin for this DSL. The use case specification language FRSL would help precisely specify the system's functional requirements and bring more automation in the software development.
{"title":"FRSL: A Domain Specific Language to Specify Functional Requirements","authors":"Duc-Hanh Dang","doi":"10.25073/2588-1086/vnucsce.803","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.803","url":null,"abstract":"In software development, to obtain a precise specification of the software system's functional requirements is significant to ensure the software quality as well as to automate the development. Use cases are an effective way to capture functional requirements, however, the use of ambiguous or vague language in the use case can lead to imprecision. It is essential to ensure that use case specifications are clear, concise, and complete to avoid imprecision in requirements. This paper aims to develop a domain specific language called FRSL to precisely specify use cases and to provide a basis for transformations to generate software artifacts from the use case specification. We define a FRSL metamodel to capture the technical domain of use cases for FRSL's abstract syntax, and then provides a textual concrete syntax for this language. We also define a formal operational semantics for FRSL by characterizing the execution of a FRSL specification as sequences of system snapshot transitions. This formal semantics on the one hand allows us to precisely explain the meaning of use cases and their relationships, on the other hand provides a basis for transformations from the use case specification. We implement a tool support for this language and perform an evaluation of its expressiveness in comparison with current use case specification languages. This work brings out (1)~a DSL to specify use cases that is defined based on a formal semantics of use cases; and (2)~a tool support realized as an Eclipse plugin for this DSL. The use case specification language FRSL would help precisely specify the system's functional requirements and bring more automation in the software development.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132392629","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}
Pub Date : 2023-04-01DOI: 10.25073/2588-1086/vnucsce.657
Duc-Hanh Dang, Thi-Hanh Nguyen
Model transformations play an essential role in model-driven engineering. However, model transformations are often complex to develop, maintain, and ensure quality. Platform-independent specification languages for transformations are required to fully and accurately express requirements of transformation systems and to offer support for realization and verification tasks. Several specification languages have been proposed, but it still lacks a strong one based on a solid formal foundation for both high expressiveness and usability. This paper introduces a language called TC4MT to precisely specify requirements of transformations. The language is designed based on a combination of a contract-based approach and the graph theory foundation of triple graph grammar. Specifically, we consider graph patterns as core elements of our language and provide a concrete syntax in the form of UML class diagrams together with OCL conditions to visually and intuitively represent such pattern-based specifications. We develop a support tool and evaluate our proposed method by comparing it with current methods in literature.
{"title":"A Contract-Based Specification Method for Model Transformations","authors":"Duc-Hanh Dang, Thi-Hanh Nguyen","doi":"10.25073/2588-1086/vnucsce.657","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.657","url":null,"abstract":"Model transformations play an essential role in model-driven engineering. However, model transformations are often complex to develop, maintain, and ensure quality. Platform-independent specification languages for transformations are required to fully and accurately express requirements of transformation systems and to offer support for realization and verification tasks. Several specification languages have been proposed, but it still lacks a strong one based on a solid formal foundation for both high expressiveness and usability. This paper introduces a language called TC4MT to precisely specify requirements of transformations. The language is designed based on a combination of a contract-based approach and the graph theory foundation of triple graph grammar. Specifically, we consider graph patterns as core elements of our language and provide a concrete syntax in the form of UML class diagrams together with OCL conditions to visually and intuitively represent such pattern-based specifications. We develop a support tool and evaluate our proposed method by comparing it with current methods in literature.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116570937","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}
Pub Date : 2022-12-16DOI: 10.25073/2588-1086/vnucsce.329
Ngan Nguyen Luu Thuy, Đặng Văn Thìn, Hoàng Xuân Vũ, Nguyễn Văn Tài, Khoa Thi-Kim Phan
Natural Language Inference (NLI) is a high-level semantic task in Natural Language Processing - NLP, and it extends further challenges if it is in the cross-lingual scenario. In recent years, pre-trained multilingual language models (e.g., mBERT ,XLM-R, InfoXLM) have greatly contributed to the success of dealing with these challenges. Based on the motivation behind these achievements, this paper describes our approach based on fine-tuning pretrained multilingual language models (XLM-R, InfoXLM) to tackle the shared task ``Vietnamese and English-Vietnamese Textual Entailment'' at the 8th International Workshop on Vietnamese Language and Speech Processing (VLSP 2021footnote{https://vlsp.org.vn/vlsp2021}). We investigate other techniques to improve the performance of our work: Cross-validation, Pseudo-labeling (PL), Learning rate adjustment, and Postagging. All experimental results demonstrated that our approach based on the InfoXLM model achieved competitive results, ranking 2nd for the task evaluation in VLSP 2021 with 0.89 in terms of F1-score on the private test set.
{"title":"vnNLI - VLSP 2021: Vietnamese and English-Vietnamese Textual Entailment Based on Pre-trained Multilingual Language Models","authors":"Ngan Nguyen Luu Thuy, Đặng Văn Thìn, Hoàng Xuân Vũ, Nguyễn Văn Tài, Khoa Thi-Kim Phan","doi":"10.25073/2588-1086/vnucsce.329","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.329","url":null,"abstract":"Natural Language Inference (NLI) is a high-level semantic task in Natural Language Processing - NLP, and it extends further challenges if it is in the cross-lingual scenario. In recent years, pre-trained multilingual language models (e.g., mBERT ,XLM-R, InfoXLM) have greatly contributed to the success of dealing with these challenges. Based on the motivation behind these achievements, this paper describes our approach based on fine-tuning pretrained multilingual language models (XLM-R, InfoXLM) to tackle the shared task ``Vietnamese and English-Vietnamese Textual Entailment'' at the 8th International Workshop on Vietnamese Language and Speech Processing (VLSP 2021footnote{https://vlsp.org.vn/vlsp2021}). We investigate other techniques to improve the performance of our work: Cross-validation, Pseudo-labeling (PL), Learning rate adjustment, and Postagging. All experimental results demonstrated that our approach based on the InfoXLM model achieved competitive results, ranking 2nd for the task evaluation in VLSP 2021 with 0.89 in terms of F1-score on the private test set.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124389717","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}
Pub Date : 2022-12-16DOI: 10.25073/2588-1086/vnucsce.375
S. Q. Nguyen, Tien Huu Vu, Duong Trieu Dinh, Minh Bao Dinh, Minh N. Do, X. Hoang
Versatile Video Coding (VVC) has been recently becoming popular in coding videos due to its compression efficiency. To reach this performance, Joint Video Experts Team (JVET) has introduced a number of improvement techniques to VVC coding model. Among them, VVC Intra coding introduces a new concept of quad-tree nested multi-type tree (QTMT) and extends the predicted modes with up to 67 options. As a result, the complexity of the VVC Intra encoding also greatly increases. To make VVC Intra coding more feasible in real-time applications, we propose in this paper a novel deep learning based fast QTMT and an early mode prediction method. At the first stage, we use a learned convolutional neural network (CNN) model to predict the coding unit map and then fed into the VVC encoder to early terminate the block partitioning process. After that, we design a statistical model to predict a list of most probable modes (MPM) for each selected Coding using (CU) size. Finally, we employ a so-called three-steps mode decision algorithm to estimate the optimal directional mode without sacrificing the compression performance. The proposed early CU splitting and fast intra prediction are integrated into the latest VTM reference software. Experimental results show that the proposed method can save 50.2% of encoding time with a negligible BD-Rate increase.
{"title":"Early CTU Termination and Three-steps Mode Decision Method for Fast Versatile Video Coding","authors":"S. Q. Nguyen, Tien Huu Vu, Duong Trieu Dinh, Minh Bao Dinh, Minh N. Do, X. Hoang","doi":"10.25073/2588-1086/vnucsce.375","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.375","url":null,"abstract":"Versatile Video Coding (VVC) has been recently becoming popular in coding videos due to its compression efficiency. To reach this performance, Joint Video Experts Team (JVET) has introduced a number of improvement techniques to VVC coding model. Among them, VVC Intra coding introduces a new concept of quad-tree nested multi-type tree (QTMT) and extends the predicted modes with up to 67 options. As a result, the complexity of the VVC Intra encoding also greatly increases. To make VVC Intra coding more feasible in real-time applications, we propose in this paper a novel deep learning based fast QTMT and an early mode prediction method. At the first stage, we use a learned convolutional neural network (CNN) model to predict the coding unit map and then fed into the VVC encoder to early terminate the block partitioning process. After that, we design a statistical model to predict a list of most probable modes (MPM) for each selected Coding using (CU) size. Finally, we employ a so-called three-steps mode decision algorithm to estimate the optimal directional mode without sacrificing the compression performance. The proposed early CU splitting and fast intra prediction are integrated into the latest VTM reference software. Experimental results show that the proposed method can save 50.2% of encoding time with a negligible BD-Rate increase.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131282797","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}
Pub Date : 2022-12-16DOI: 10.25073/2588-1086/vnucsce.330
Thin Dang Van, D. Hao, N. Nguyen, Luân Đình Ngô, Kiến Lê Hiếu Ngô
In this paper, we describe an empirical study of data augmentation techniques with various pre-trained language models on the bilingual dataset which was presented at the VLSP 2021 - Vietnamese and English-Vietnamese Textual Entailment. We apply the machine translation tool to generate new training set from original training data and then investigate and compare the effectiveness of a monolingual and multilingual model on the new data set. Our experimental results show that fine-tuning a pre-trained multilingual language XLM-R model with an augmented training set gives the best performance. Our system was ranked third in the shared-task VLSP 2021 with the F1-score of about 0.88.
{"title":"vnNLI - VLSP2021: An Empirical Study on Vietnamese-English Natural Language Inference Based on Pretrained Language Models with Data Augmentation","authors":"Thin Dang Van, D. Hao, N. Nguyen, Luân Đình Ngô, Kiến Lê Hiếu Ngô","doi":"10.25073/2588-1086/vnucsce.330","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.330","url":null,"abstract":"In this paper, we describe an empirical study of data augmentation techniques with various pre-trained language models on the bilingual dataset which was presented at the VLSP 2021 - Vietnamese and English-Vietnamese Textual Entailment. We apply the machine translation tool to generate new training set from original training data and then investigate and compare the effectiveness of a monolingual and multilingual model on the new data set. Our experimental results show that fine-tuning a pre-trained multilingual language XLM-R model with an augmented training set gives the best performance. Our system was ranked third in the shared-task VLSP 2021 with the F1-score of about 0.88.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126472102","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}
Pub Date : 2022-12-16DOI: 10.25073/2588-1086/vnucsce.354
Tran Nguyen Huong
In recent years, automated test data generation from source code has gained a significantpopularity in software testing. This paper proposes a method, named Hybrid, to generate test datafor unit testing C/C++ projects. The method is a combination of two test data generation methodsnamed IBVGT and WCFT. In IBVGT method, the source code is analyzed to find simple conditions.Then, bases on these conditions, IBVGT generates test data for boundary values without having tosolve test paths constraints. This makes the method faster than BVTG method when generating testdata. In Hybrid method, while generating test data using WCFT, simple conditions are collected forboundary values test data generation. Test data generated by Hybrid are able to ensure both highsource code coverage and error detection ability. In addition, Hybrid is capable of finding infeasibleexecution paths and dead code. Experimental results with some popular unit functions show thatHybrid outperforms STCFG method in terms of test data generation time and boundary values relatederror detection. IBVGT is superior to BVTG in term of test data generation time whilst its boundaryvalues related error detection ability depends on the number of simple conditions inside each unitfunction.
{"title":"A Hybrid Method for Test Data Generation for Unit Testing of C/C++ Projects","authors":"Tran Nguyen Huong","doi":"10.25073/2588-1086/vnucsce.354","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.354","url":null,"abstract":"In recent years, automated test data generation from source code has gained a significantpopularity in software testing. This paper proposes a method, named Hybrid, to generate test datafor unit testing C/C++ projects. The method is a combination of two test data generation methodsnamed IBVGT and WCFT. In IBVGT method, the source code is analyzed to find simple conditions.Then, bases on these conditions, IBVGT generates test data for boundary values without having tosolve test paths constraints. This makes the method faster than BVTG method when generating testdata. In Hybrid method, while generating test data using WCFT, simple conditions are collected forboundary values test data generation. Test data generated by Hybrid are able to ensure both highsource code coverage and error detection ability. In addition, Hybrid is capable of finding infeasibleexecution paths and dead code. Experimental results with some popular unit functions show thatHybrid outperforms STCFG method in terms of test data generation time and boundary values relatederror detection. IBVGT is superior to BVTG in term of test data generation time whilst its boundaryvalues related error detection ability depends on the number of simple conditions inside each unitfunction.","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129205211","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}
Pub Date : 2022-12-16DOI: 10.25073/2588-1086/vnucsce.371
Doanh Bui Cao, Truc Thi Thanh Trinh, Trong-Thuan Nguyen, V. Nguyen, Nguyen D. Vo
The automatic image caption generation is attractive to both Computer Vision and Natural Language Processing research community because it lies in the gap between these two fields. Within the vieCap4H contest organized by VLSP 2021, we participate and present a Transformer-based solution for image captioning in the healthcare domain. In detail, we use grid features as visual presentation and pre-training a BERT-based language model from PhoBERT-base pre-trained model to obtain language presentation used in the Adaptive Decoder module in the RSTNet model. Besides, we indicate a suitable schedule with the self-critical training sequence (SCST) technique to achieve the best results. Through experiments, we achieve an average of 30.3% BLEU score on the public-test round and 28.9% on the private-test round, which ranks 3rd and 4th, respectively. Source code is available at https://github.com/caodoanh2001/uit-vlsp-viecap4h-solution.
{"title":"vieCap4H Challenge 2021: A transformer-based method for Healthcare Image Captioning in Vietnamese","authors":"Doanh Bui Cao, Truc Thi Thanh Trinh, Trong-Thuan Nguyen, V. Nguyen, Nguyen D. Vo","doi":"10.25073/2588-1086/vnucsce.371","DOIUrl":"https://doi.org/10.25073/2588-1086/vnucsce.371","url":null,"abstract":"The automatic image caption generation is attractive to both Computer Vision and Natural Language Processing research community because it lies in the gap between these two fields. Within the vieCap4H contest organized by VLSP 2021, we participate and present a Transformer-based solution for image captioning in the healthcare domain. In detail, we use grid features as visual presentation and pre-training a BERT-based language model from PhoBERT-base pre-trained model to obtain language presentation used in the Adaptive Decoder module in the RSTNet model. Besides, we indicate a suitable schedule with the self-critical training sequence (SCST) technique to achieve the best results. Through experiments, we achieve an average of 30.3% BLEU score on the public-test round and 28.9% on the private-test round, which ranks 3rd and 4th, respectively. Source code is available at https://github.com/caodoanh2001/uit-vlsp-viecap4h-solution. \u0000 ","PeriodicalId":416488,"journal":{"name":"VNU Journal of Science: Computer Science and Communication Engineering","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127272648","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：