Vol. 47, No. 11, Nov. 2020
Digital Library


Binary-Compatible User-Mode Polling-Based Inter-VM Communication Techniques Using Shared Memory

Jihong Min, Juhyung Park, Joonseok Park

http://doi.org/10.5626/JOK.2020.47.11.1015

Virtual machines commonly use TCP/IP protocol to exchange data with the host or other virtual machines, but the protocol is inefficient. Communication using inter-VM shared memory can be used for better efficiency, but a disadvantage forces existing TCP/IP-based programs to be reprogrammed or recompiled. There have been several studies on the binary-compatible inter-VM shared memory-based communication methods to resolve this issue, yet the overheads exist. In this paper, we propose techniques that reduce the overheads of the current binary-compatible inter-VM shared memory-based communication methods. Our scheme bypasses the existing network stack with function hooking to TCP/IP library, and introduces the transmission queue per connection and user-mode polling technique to remove the kernel-mode switching overhead. The experiment results show that the latency can be reduced by 96.96% and throughtput can be increased by 222.24% on average by using the proposed techniques compared to the existing virtual network.

Korean Semantic Role Labeling with BERT

Jangseong Bae, Changki Lee, Soojong Lim, Hyunki Kim

http://doi.org/10.5626/JOK.2020.47.11.1021

Semantic role labeling is an application of natural language processing to identify relationships such as "who, what, how and why" with in a sentence. The semantic role labeling study mainly uses machine learning algorithms and the end-to-end method that excludes feature information. Recently, a language model called BERT (Bidirectional Encoder Representations from Transformers) has emerged in the natural language processing field, performing better than the state-of- the-art models in the natural language processing field. The performance of the semantic role labeling study using the end-to-end method is mainly influenced by the structure of the machine learning model or the pre-trained language model. Thus, in this paper, we apply BERT to the Korean semantic role labeling to improve the Korean semantic role labeling performance. As a result, the performance of the Korean semantic role labeling model using BERT is 85.77%, which is better than the existing Korean semantic role labeling model.

Low-Resolution Image Classification Using Knowledge Distillation From High-Resolution Image Via Self-Attention Map

Sungho Shin, Joosoon Lee, Junseok Lee, Seungjun Choi, Kyoobin Lee

http://doi.org/10.5626/JOK.2020.47.11.1027

Traditional deep-learning models have been developed using high-quality images. However, when the low resolution images are rendered, the performances of the model drop drastically. To develop a deep-learning model that can respond effectively to low-resolution images, we extracted the information from the model, which uses high-resolution images as input, in the form of the Attention Map. Using the knowledge distillation technique, the information delivering Attention Map, extracted from the high-resolution images to low-resolution image models, could reduce the error rate by 2.94%, when classifying the low-resolution CIFAR images of 16×16 resolution. This was at 38.43% of the error reduction rate when the image resolution was lowered from 32×32 to 16×16, which could demonstrate excellence in this network.

Detection of Hate Speech with Emotes in Online Streaming Chat: Based on Deep Learning Model

Jaeheon Kim, Donghee Yvette Wohn, Meeyoung Cha

http://doi.org/10.5626/JOK.2020.47.11.1032

Through the development of the natural language processing, artificial intelligence for the detection of hate speech has also grown. However, malicious users continuously produce new forms of hate speech not identified by the machine. Of the various types is a mixture of text and emotes, easily distinguishable by humans but not by rule-based detection. This study analyzes chat data of Twitch.tv, a popular online streaming service, to identify types of a mixture of text and emotes. We compared the emotes usage in the hate speech and the race of the streamer. Besides, a new method using the bidirectional long short-term memory model is proposed to detect new hate speech types. Among approximately 15 million chats, the proposed method could identify additional hate speech with an F1-score of 0.745.

English-Korean Neural Machine Translation using MASS with Relative Position Representation

Youngjun Jung, Cheoneum Park, Changki Lee, Junseok Kim

http://doi.org/10.5626/JOK.2020.47.11.1038

Neural Machine Translation has been mainly studied for a Sequence-to-Sequence model using supervised learning. However, since the supervised learning method shows low performance when the data is insufficient, recently, a transfer learning method of fine-tuning using the pre-training model based on a large amount of monolingual data such as BERT and MASS has been mainly studied in the field of natural language processing. In this paper, MASS using the pre-training method for language generation, was applied to the English-Korean machine translation. As a result of the experiment, the performance of the English-Korean machine translation model using MASS showed better performance than the existing models, and the performance of the machine translation model was further improved by applying the relative position representation method to MASS.

Behavior Model-Based Fault Localization for RESTful Web Applications

Jong-In Jang, Nakwon Lee, Duksan Ryu, Jongmoon Baik

http://doi.org/10.5626/JOK.2020.47.11.1044

Because of the nature of Web applications being more complex, larger in scale and more likely to be composed of black box components compared to traditional software systems wherein fault localization techniques are actively used, existing techniques can be only minimally applied to localize faults in Web applications. Also, existing studies to localize a fault in a complex system such as a Web application system also have limitations in capturing the indirect interactions in Web applications and suffers from the Web application’s dynamic nature. In this study, we propose a behavior modeling-based fault localization for the RESTful Web applications. The approach models a RESTful Web application as a sequence of behaviors that captures the direct and indirect interactions in the application. The modeling process is lightweight and it is not necessary to build the model in advance of the actual execution of application. The spectrum-based fault localization is then performed in the granularity of behavior pairs in the behavior model. To demonstrate the approach, a case study on the RESTful Web application built upon the YouTube Data API v3 was conducted and demonstrated that the approach can successfully resolve aforementioned difficulties and localize a fault in the RESTful Web application.

Deep Learning-based Text Classification Model for Poisonous Clauses Classification

Gihyeon Choi, Youngjin Jang, Harksoo Kim, Kwanwoo Kim

http://doi.org/10.5626/JOK.2020.47.11.1054

Most companies sign contracts based on the contract prior to executing the task. However, several problems can occur if the poisonous clauses are not identified before the contract is concluded. To prevent this problem, companies have an expert review the contract, but the service requires much time and money. If there is a system in which the poisonous clauses can be identified through prior review of the contract, the high cost and time incurred in reviewing the contract can be mitigated. Thus, this paper proposes a text classification model that identifies any poisonous clause in the contract by inputing each paragraph in the contract. To improve the classification performance of the proposed model, the importance of each sentence is calculated based on the relationship information between the sentence in the paragraph and the class to be classified, and classification is performed by reflecting it in each sentence. The proposed model showed the performance of the F1 score 84.51%p in experiments using actual contract data and the highest performance with the F1 score 93.64%p in experiments using the WOS-5736 dataset for the performance comparison with the existing text classification models.

LEXAI : Legal Document Similarity Analysis Service using Explainable AI

Juho Bai, Seog Park

http://doi.org/10.5626/JOK.2020.47.11.1061

Recently, in keeping with the improvement of deep learning, studies on using deep learning a specialized field have diversified. Semantic searching for legal documents is an essential part of the legal field. However, it is difficult to function outside of the service using the expert system because it requires professional knowledge in the relevant field. It is also challenging to establish an automated, semantically similar legal document retrieval environment because the cost of hiring professional human resources is high. While existing retrieval services provide an environment based on expert systems and statistical systems, the proposed method adopts the deep learning method with a classification task. We propose a database system structure that provides searching for legal documents with high semantic similarity using an explainable neural network. The features of these proposed methods show the performance of developing and verifying visual similarity assessment methods for semantic relevance among similar documents.

Formal Model Design for Network and Operating System Behaviors in Real-time Distributed System Verification with Coq

Yoonseung Kim, Chung-Kil Hur

http://doi.org/10.5626/JOK.2020.47.11.1071

Improving the safety and reliability of distributed systems using formal verification methods is an urgent problem. As many of these distributed systems are safety-critical, such as medical or avionics systems, failures of these systems may cause catastrophic results. However, applying formal verification to distributed systems requires not only execution semantics in software, but also behavioral models of the environments, including the operating systems and network involved. We designed a formal abstract model of network and operating system behaviors with the Coq proof assistant. This model consists of local-site execution semantics that model a single computer, the composition of these local-site semantics along with a message exchange model constitutes the global system semantics. We applied and tested this model to verify its applicability when used in a simple server-client system. We expect this model to be used in the verification of practical systems.

Performance Comparison and Analysis Between Neural and Non-neural Autoencoder-based Recommender Systems

Yoonki Jeong, Jongwuk Lee

http://doi.org/10.5626/JOK.2020.47.11.1078

While deep neural networks have been bringing advances in many domains, recent studies have shown that the performance gain from deep neural networks is not as extensive as reported, compared to the higher computational complexity they require. This phenomenon is caused by the lack of shared experimental settings and strict analysis of proposed methods. In this paper, 1) we build experimental settings for fair comparison between the different recommenders, 2) provide empirical studies on the performance of the autoencoder-based recommender, which is one of the main families in the literature, and 3) analyze the performance of a model according to user and item popularity. With extensive experiments, we found that there was no consistent improvement between the neural and the non-neural models in every dataset and there is no evidence that the non-neural models have been improving over time. Also, the non-neural models achieved better performance on popular item accuracy, while the neural models relatively perform better on less popular items.

Data Modelling Method for Real-Time Advertising Service Based on Viewer Reaction and Intention in Online Broadcasting

Seongju Kang, Chaeeun Jeong, Kwangsue Chung

http://doi.org/10.5626/JOK.2020.47.11.1086

The interaction between the existing advertising service and the user is limited. To provide a personalized advertising service, advertisement systems should predict the user"s preference based on the user"s profile and the user-content relationship. Many recommendation schemes have been studied to predict the preferences of users. However, the existing recommendation system is difficult to guarantee real-time preference prediction as it performs a calculation of the matrix with high computational complexity. In this paper, we propose a data modeling method for real-time advertising services based on the reaction and intention of viewers. To predict the user"s preference in real-time, the user"s historical data is modeled in a tree structure. The tree structure allows us to retrieve and compare the data with logarithmic time complexity. To improve the accuracy of the recommendation, we have proposed a recommendation algorithm that considers both the user"s positive and negative evaluations. Finally, we have evaluated the performance of the proposed method through various methods.

Energy-efficient Video Streaming System with TCP/UDP over Heterogeneous Wireless Networks

Yunmin Go, Hyunmin Noh, Jeung Won Choi, Kyungwoo Kim, Kihun Kim, Jongman Lee, Hyun Park, Hwangjun Song

http://doi.org/10.5626/JOK.2020.47.11.1092

In this paper, we propose a video streaming system with TCP/UDP over heterogeneous wireless networks. Unlike the existing HTTP adaptive streaming system, the proposed system can use both TCP and UDP to overcome the TCP"s performance degradation problem and improve energy efficiency. The proposed system determines the type of transport protocol, the video quality, and the amount of data received through each wireless network according to the network conditions, buffer occupancy, and energy efficiency of the mobile device. When UDP is selected as the transport protocol, the proposed system employs raptor code to provide more reliable data transmission. And raptor parameters, such as symbol size, the amount of redundant data, are also determined to provide energy-efficient transmission. The proposed system is implemented in an android device and verified in real WiFi and LTE networks. Also, it is verified that the proposed system can provide high energy-efficient and high-quality streaming services.


Search




Journal of KIISE

  • ISSN : 2383-630X(Print)
  • ISSN : 2383-6296(Electronic)
  • KCI Accredited Journal

Editorial Office

  • Tel. +82-2-588-9240
  • Fax. +82-2-521-1352
  • E-mail. chwoo@kiise.or.kr