Vol. 45, No. 9,
Sep. 2018
Digital Library
Vulnerability Analysis on Kernel Code and Memory Protection in Nested Kernel
http://doi.org/10.5626/JOK.2018.45.9.873
Nested Kernel is a secure kernel architecture, presented at the 2015 ACM ASPLOS conference, which aims at assuring the lifetime integrity of the kernel. With the conventional off-the-shelf HW-based protection facility, the Nested Kernel significantly improves the security of the system by introducing a new OS kernel architecture. However, our analysis reveals that the current Nested Kernel has some flaws in its implementation for handling direct mapping and the kernel code mapping region. In addition, its integrity can be broken because of the reported security vulnerability. Consequently, the Nested Kernel needs further study for it to be used safely as a security kernel.
Real-time Intrinsic Image Decomposition using 3D Geometry of Reconstructed Indoor Scene
http://doi.org/10.5626/JOK.2018.45.9.881
Development of 3D reconstruction techniques has improved the accessibility of image-based modeling. However, the usage of reconstructed models is limited because of shading in texture. Therefore, we present an intrinsic image decomposition method for indoor models reconstructed using commercial software to manipulate light conditions dynamically. In our approach, we use light source estimation and inverse rendering to perform intrinsic image decomposition relative to reconstructed geometry information. We apply Phong reflectance model and light propagation volume, real-time rendering method, to inverse rendering for real-time performance. Considering the features of indoor scenes, all light sources are estimated as point lights and indirect lighting include reflectance colors estimated from filtered texture. Our approach maintains simplification of lighting, benefit of image-based modeling, because of estimating reflectance relative to geometric information.
User-Weighted Viewpoint/Lighting Control for Multi-Object Scene
http://doi.org/10.5626/JOK.2018.45.9.888
In computer graphics, viewpoint selection for objects in a scene has been performed by evaluating the goodness of sampled viewpoints. Since the definition of a good viewpoint varies according to the user’s purpose, various measurements such as entropy and mesh saliency have been used. In this paper, we propose a method of selecting the best viewpoint and lighting for a multi-object scene, based on the user-assigned importance of each object. After sampling a viewpoint and lighting from the surrounding sphere of the scene, we render the images by combining the sampled viewpoint and lighting. We then select the best result that coincides with user-assigned importance by quantifying the saliency of each object in the rendered image. While this technique has the disadvantage of high computation cost due to the need to render combinations of viewpoints and lighting, it obtains the viewpoint and lighting most suitable for the user`s needs. In order to minimize the computation cost, an object-by-object pixel classification technique on GPU is also proposed in this paper.
Knowledge Completion Modeling using Knowledge Base Embedding
Hyun-Young Choi, Ji-Hun Hong, Wan-Gon Lee, Batselem Jagvaral, Myung-Joong Jeon, Hyun-Kyu Park, Young-Tack Park
http://doi.org/10.5626/JOK.2018.45.9.895
In recent years, a number of studies have been conducted for the purpose of automatically building a knowledge base that is based on web data. However, due to the incomplete nature of web data, there can be missing data or a lack of connections among the data entities that are present. In order to solve this problem, recent studies have proposed methods that train a model to predict this missing data through an artificial neural network based on natural language embedding, but there is a drawback to embedding entities. In practice, natural language corpus is not present in many knowledge bases. Therefore, in this paper, we propose a knowledge completion method that converts the knowledge base of RDF data into an RDF-sentence and uses embedding to create word vectors. We conducted a triple classification experiment in order to measure the performance of the proposed method. The proposed method was then compared with existing NTN models, and on average, 15% accuracy was obtained. In addition, we obtained 88%accuracy by applying the proposed method to the Korean knowledge base known as WiseKB.
A Practice of Software Development Process Visualization for Army Information System Management
Woo Sung Jang, Hyung Seung Son, R.Young Chul Kim, Jong Hoon Lee
http://doi.org/10.5626/JOK.2018.45.9.904
To increase the chance of success of the current software project, we need a software process that is like a manufacturing process. Specially, the army information system should focus on developing an effective information system and managing byproducts as a whole process. Moreover, maintenance will require a way to avoid maintenance delays, which increase a costs, and degrade quality. To solve this problem, we apply a process visualization mechanism for the Army’s system, one of process, architecture, and documentation visualization in NIPA. We can guarantee a high quality of software that will provide transparency and traceability for all stakeholders in the life cycle. We expect to maintain high quality for the army’s software with quality indicators, result visualization, scheduling control, and management.
Speech-Act Analysis System Based on Dialogue Level RNN-CNN Effective on the Exposure Bias Problem
http://doi.org/10.5626/JOK.2018.45.9.911
The speech-act is the intention of the speaker in his or her utterance. Speech-act analysis classifies the speech-act about a given utterance. Recently, a lot of research based on machine learning using a corpus have been done. We have two goals in this study. First, the utterances in dialogue are continuative and organically related to each other, and the speech-act of a current utterance is greatly influenced by the direct previous utterance. Second, previous research did not deal with the exposure bias problem when the speech-act analysis model use the speech-act result of a previous utterance. In this paper, we suggest the RNN-CNN dialogue-level speech-act analysis model. We also experiment with the exposure bias problem. Finally, the RNN-CNN shows an 86.87% performance on the oracle condition and an 86.27% performance on the greedy condition.
Knowledge Base Population Model Using Non-Negative Matrix Factorization
Jiho Kim, Sangha Nam, Key-Sun Choi
http://doi.org/10.5626/JOK.2018.45.9.918
The purpose of a knowledge base is to incorporate all the knowledge in the world in a format that machines can understand. In order for a knowledge base to be useful, it must continuously acquire and add new knowledge. However, it cannot if it lacks knowledge-acquisition ability. Knowledge is mainly acquired by analyzing natural language sentences. However, studies on internal knowledge acquisition are being neglected. In this paper, we introduce a non-negative matrix factorization method for knowledge base population. The model introduced in this paper transforms a knowledge base into a matrix and then learns the latent feature vector of each entity tuple and relation by decomposing the matrix and reassembling the vectors to score the reliability of the new knowledge. In order to demonstrate the effectiveness and superiority of our method, we present results of experiments and analysis performed with Korean DBpedia.
Resolution of Answer-Repetition Problems in a Generative Question-Answering Chat System
http://doi.org/10.5626/JOK.2018.45.9.925
A question-answering (QA) chat system is a chatbot that responds to simple factoid questions by retrieving information from knowledge bases. Recently, many chat systems based on sequence-to-sequence neural networks have been implemented and have shown new possibilities for generative models. However, the generative chat systems have word repetition problems, in that the same words in a response are repeatedly generated. A QA chat system also has similar problems, in that the same answer expressions frequently appear for a given question and are repeatedly generated. To resolve this answer-repetition problem, we propose a new sequence-to-sequence model reflecting a coverage mechanism and an adaptive control of attention (ACA) mechanism in a decoder. In addition, we propose a repetition loss function reflecting the number of unique words in a response. In the experiments, the proposed model performed better than various baseline models on all metrics, such as accuracy, BLEU, ROUGE-1, ROUGE-2, ROUGE-L, and Distinct-1.
Korean Machine Reading Comprehension using Reinforcement Learning and Dual Co-Attention Mechanism
http://doi.org/10.5626/JOK.2018.45.9.932
Machine Reading Comprehension is a question-answering model for the purposes of understanding a given document and then finding the correct answer within the document. Previous studies on the Machine Reading Comprehension model have been based on end-to-end neural network models with various attention mechanisms. However, in the previous models, difficulties arose when attempting to find answers with long dependencies between lexical clues because these models did not use grammatical and syntactic information. To resolve this problem, we propose a Machine Reading Comprehension model with a dual co-attention mechanism reflecting part-of-speech information and shortest dependency path information. In addition, to increase the performances, we propose a reinforce learning method using F1-scores of answer extraction as rewards. In the experiments with 18,863 question-answering pairs, the proposed model showed higher performances (exact match: 0.4566, F1-score: 0.7290) than the representative previous model.
Automatic Segmentation of Renal Parenchyma using Shape and Intensity Information based on Multi-atlas in Abdominal CT Images
Hyeonjin Kim, Helen Hong, Kidon Chang, Koon Ho Rha
http://doi.org/10.5626/JOK.2018.45.9.937
Renal parenchyma segmentation is necessary to predict contralateral hypertrophy after renal partial nephrectomy. In this paper, we propose an automatic segmentation method of renal parenchyma using shape and intensity information based on the multi-atlas in abdominal CT images. First, similar atlases are selected using volume-based similarity registration and intensity-similarity measure. Second, renal parenchyma is segmented using two-stage registration and constrained intensity-based locally-weighted voting. Finally, renal parenchyma is refined using a Gaussian mixture model-based multi-thresholds and shape-prediction map in under- and over-segmented data. The average dice similarity coefficient of renal parenchyma was 91.34%, which was 18.19%, 1.35% higher than the segmentation method using majority voting and locally-weighted voting in dice similarity coefficient, respectively.
Word Embedding using Relative Position Information between Words
Hyunsun Hwang, Changki Lee, HyunKi Jang, Dongho Kang
http://doi.org/10.5626/JOK.2018.45.9.943
In Word embedding, which is used to apply deep learning to natural language processing, a word is expressed on a vector space. This has the advantage of dimension reduction, whereby similar words have similar vector values. Word embedding needs to learn large-scale corpus to get achieve good performance. However, the word2vec model, which has frequently been used in the past, has a disadvantage in that it does not use relative position information between words because it largely learns the word appearance rate by simplifying the model for large capacity corpus learning. In this paper, we modified the existing word embedding learning model to enable it to learn using relative position information between words. Experimental results show that the performance of the word-analogy of the proposed modified word embedding learning model is improved when word embedding is learned using relative position information between words.
Multi-Level Fusion Activity Recognition Framework using Smart Devices
http://doi.org/10.5626/JOK.2018.45.9.950
Traditional inertial sensor based activity recognition methods in which multiple sensor units are attached to the body is changing to accommodate the use of smart devices such as smartphones and smartwatches. In this paper, we propose a multi-level fusion activity recognition framework to recognize daily activities using smartphones and smartwatches which can be purchased easily for minimum sensor based activity recognition. The proposed framework uses various types of fusion techniques such as data fusion, feature fusion, and decision fusion. While the proposed framework does not use common methods of decision fusion such as majority voting or weighted voting, it does use posterior probability based fusion for better accuracy and confidence. Experiments are conducted to compare results between using and not using the probability and between using and not using each fusion technique. The results demonstrated the excellent performance of the proposed framework.
Privacy Budget Allocation Technique Based on Variable Length Window for Traffic Data Publishing with Differential Privacy in Road Networks
Gunhyung Jo, Kangsoo Jung, Seog Park
http://doi.org/10.5626/JOK.2018.45.9.957
Recently, traffic volume data at every timestamp have been required in many fields such as road design and traffic analysis. Such traffic volume data may contain individual sensitive location information, which leads to privacy violation such as personal route exposure. Differential privacy has the advantage of protecting sensitive personal information in this situation while controlling the data utility by inserting noise to raw data. However, because of the traffic volume data generally would be an infinite size over time, there is a drawback in that data is useless because insufficiently large scaled noise is inserted. In order to overcome this drawback, researches have been conducted on applying the differential privacy technique only to the traffic volume data contained in windows of a certain time range. However, in the previous studies, the length of the window was fixed, inducing a limit whereby the correlation of the road sections and the time-specificity are not considered. In this paper, we propose a variable length window technique considering the correlation between road segments and time-specificity.
Priority-based Multi-level MQTT System to Provide Differentiated IoT Services
Geonwoo Kim, Jiwoo Park, Kwangsue Chung
http://doi.org/10.5626/JOK.2018.45.9.969
MQTT (Message Queue Telemetry Transport) is a typical lightweight protocol that uses the Publish/Subscribe method to send messages. It also provides three levels of QoS (Quality of Service) to ensure reliable delivery of messages. However, since MQTT does not support priority processing, it cannot provide a differentiated service in an environment requiring rapid processing, such as an emergency medical center. In this paper, we propose a system that processes messages according to priority by adding priority flags to fixed headers of MQTT messages. The proposed system uses multi-level queues to provide priority-based differentiated services and guarantees minimal message throughput. Through experiments, we have confirmed that the proposed system reduces the end-to-end delay according to the priorities and ensures minimum message throughput in simulated environments where many messages are transmitted and delay occurs.
Case Study for Collecting Policy Evaluation Factors upon Request when Creating XACML Policy
http://doi.org/10.5626/JOK.2018.45.9.975
As the Internet of Things environment continues to expand, access control issues continue to emerge. OneM2M, one of the standards of the IOT platform, allows access control using XACML. In the arena of access control, conflicts must be solved. Because of this, various solutions are being investigated in order to solve these problems. Currently, however, the policy editor must solve policy conflicts by themselves. So, the policy editor needs to be able to effectively collect information about policies and conditions that affect their policy evaluation decisions in order to resolve policy conflicts. In this paper, we analyze policy and express policy evaluation methods in terms of Truth Table. In addition, we present a tree-based policy evaluation factor collection method through a case study on a policy evaluation factor collection method according to requests using Truth Table.
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