Vol. 46, No. 8,
Aug. 2019
Digital Library
Deep Learning Model based on Autoencoder for Reducing Algorithmic Bias of Gender
http://doi.org/10.5626/JOK.2019.46.8.721
Algorithmic bias is a discrimination that is reflected in the model by a bias in data or combination of characteristics of model and data in the algorithm. In recent years, it has been identified that the bias is not only present but also amplified in the deep learning model; thus, there exists a problem related to bias elimination. In this paper, we analyze the bias of the algorithm by gender in terms of bias-variance dilemma and identify the cause of bias. To solve this problem, we propose a deep auto-encoder based latent space matching model. Based on the experimental results, it is apparent that the algorithm bias in deep learning is caused by difference of the latent space for each protected feature in the feature extraction part of the model. A model proposed in this paper achieves the low bias by reducing the differences in extracted features by transferring data with different gender characteristics to the same latent space. We employed Equality of Odds and Equality of Opportunity as a quantitative measure and proved that proposed model is less biased than the previous model. The ROC curve shows a decrease in the deviation of the predicted values between the genders.
Improving Performance of Flash Storage Using Restricted Copyback
Duwon Hong, Seulgi Shin, Jihong Kim
http://doi.org/10.5626/JOK.2019.46.8.726
In case of modern flash-based SSDs, the performance overhead of internal data migrations is dominated by the data transfer time and not by the flash program time as in old SSDs. In order to mitigate the performance impact of data migrations, we propose rcopyback, a restricted version of copyback. Rcopyback works in a manner similar to the original copyback except that only n consecutive copybacks are allowed. By limiting the number of successive copybacks, the version guarantees internal migration of data using rcopyback without any reliability problem. In order to take a full advantage of rcopyback, we developed a rcopyback-aware FTL, rcFTL, which intelligently decides whether rcopyback should be used or not by exploiting varying host workloads. Our evaluation results show that rcFTL can improve the overall I/O throughput by 54% on average over an existing FTL which does not use copybacks.
Beyond Head-mounted Display: Extended Field of View using Sparse Peripheral Display Techniques
http://doi.org/10.5626/JOK.2019.46.8.732
The field of view (FoV) is one of the main properties of virtual reality (VR) systems with head-mounted displays (HMDs). While preceding VR studies have suggested novel methodologies to extend the FoV of HMDs, there were limitations in those methods, including: heavy weight, high cost, and screen distortion. The observation through the human retina indicated that the density of human photoreceptors are distributed variedly between the central and peripheral visual fields. The sparse peripheral display (SPD) was proposed based on these observations. In this study, we suggest and compare the combination of SPD technologies using different number (2-80) of light-emitting diodes (LEDs) and different tasks (visual search and emotion tasks). The results showed that the visual-search task with use of six LEDs and there were no significant differences in the emotion task. Discussion of the potential applications of SPD-HMD systems, including therapeutic techniques to assist handicapped people was done.
Safety Activity Management Process and a Method for Implementation of Support System based on a Multi-view System Modeling
http://doi.org/10.5626/JOK.2019.46.8.741
To ensure safety in a safety critical system, it is of importance to identify the potential hazards, to analyze and evaluate them from a concept phase. If the hazards are assessed as critical, they must be removed or safety requirements to mitigate them should be defined. Also, the safety requirements should be designed, implemented, and verified through testing. For the developers and safety experts to work together in the lifecycle, it is necessitated that the safety activities should be integrated into the modeling of system development. For the appropriate performance of these activities, a management system is required that can provide the traceability and status accounting of artifacts such as requirements specification, hazard list and safety requirements as mitigation method, design, implementation, and test cases. However, previous researches have mostly focused on providing better modeling and analysis methods for safety activities. Not much effort has been made on the management process of safety activities. In this paper, we propose the framework for safety activity management process and SLMS (Safety Lifecycle Management System) based on the multi-system modeling process, SAMM (Safety Analysis based on Multi-view Modeling). To demonstrate the capability of the proposed framework, an appropriate example of applying it to a radiotherapy system is also presented.
Learning Semantic Features for Dense Video Captioning
http://doi.org/10.5626/JOK.2019.46.8.753
In this paper, we propose a new deep neural network model for dense video captioning. Dense video captioning is an emerging task that aims at both localizing and describing all events in a video. Unlike many existing models, which use only visual features extracted from the given video through a sort of convolutional neural network(CNN), our proposed model makes additional use of high-level semantic features that describe important event components such as actions, people, objects, and backgrounds. The proposed model localizes temporal regions of events by using LSTM, a recurrent neural network(RNN). Furthermore, our model adopts an attention mechanism for caption generation to selectively focus on input features depending on their importance. By conducting experiments using a large-scale benchmark dataset for dense video captioning, AcitivityNet Captions, we demonstrate high performance and superiority of our model.
Improving Reliability of Smart Contracts and DApps by Applying Property-based and Model-based Test Methods to Different Test Levels
Kyeongsic Min, Jung-Won Lee, Byungjeong Lee
http://doi.org/10.5626/JOK.2019.46.8.763
Smart contract technology based on the blockchain enables transparent transactions and automated contract execution without third-party intervention. Ethereum provides Solidity and EVM (Ethereum Virtual Machine) that can be used to implement smart contracts. In addition, it can be used to create a DApp (Decentralized Application) without developing a new blockchain using smart contract. However, the source codes cannot be updated in smart contracts. Therefore, a lot of work is needed to fix even minor errors following deployment. Therefore, the source code should be thoroughly tested or analyzed prior to the deployment of the contract to ensure that it is free of defects. In this paper, we propose a method to identify the errors and verify the accuracy of smart contracts and DApps using dynamic testing methods. Toward this end, we defined the dynamic model needed in each test level and applied the current testing methodology, using property-based and model-based testing.
OSDEF: An Integrated Framework for Practicing Object-Oriented Software Development
http://doi.org/10.5626/JOK.2019.46.8.774
Software development starts with a specific software development process (SDP) which contains the start and end of the development, SDP plays an important role in the software engineering education. Object-oriented software development education uses several examples that contain object-oriented characteristics into education and practices. However, there is an immense burden on the implantation phases as per the scale of the program, thus creating difficulties in the identification of the connected relations between software design and implementation. In the present work, we propose the OSDEF (Object-oriented based Software Development Education Framework) framework for proceeding an efficient software engineering education based on the OOPT (Object-Oriented Process with Traceability). The framework contains artifact management tools which can directly write development artifacts inside the tool, traceability analysis tool, and emulating environment for embedded software, which can practice a layered architecture in an efficient manner.
Identifying Causes of an Accident in STPA Using the Scenario Table
http://doi.org/10.5626/JOK.2019.46.8.787
In recent years, the complexity of safety-critical systems has increased, along with the importance of the software. The software, which has become the control center of the safety system, generates control actions to control the system and then repeats the interaction of controls that re-enters the feedback generated. STPA (System Theoretic Process Analysis) is one of the hazard analysis techniques, and it analyzes the system from the perspective of the interaction of control then uses accident scenarios to identify and analyze the cause of unsafe control actions to derive safe requirements. In order to minimize omissions in the identification stage of STPA accident scenarios associated with safety requirements, in this paper we describe how to incorporate commonalities and complement vulnerabilities in the approaches described in previous studies. To do this, we propose the detailed procedure for identifying accident scenarios and the scenario table to assist them. The ultimately proposed scenario table is identified by applying it to the hazard analysis of the railway diorama system.
An Automatic Method of Generating a Large-Scale Train Set for Bi-LSTM based Sentiment Analysis
http://doi.org/10.5626/JOK.2019.46.8.800
Sentiment analysis using deep learning requires a large-scale train set labeled sentiment. However, direct labeling of sentiment by humans is time and cost-constrained, and it is not easy to collect the required data for sentiment analysis from many data. In the present work, to solve the existing problems, the existing sentiment lexicon was used to assign sentiment score, and when there was sentiment transformation element, the sentiment score was reset through dependency parsing and morphological analysis for automatic generation of large-scale train set labeled with the sentiment. The Top-k data with high sentiment score was extracted. Sentiment transformation elements include sentiment reversal, sentiment activation, and sentiment deactivation. Our experimental results reveal the generation of a large-scale train set in a shorter time than manual labeling and improvement in the performance of deep learning with an increase in the amount of train set. The accuracy of the model using only sentiment lexicon was 80.17% and the accuracy of the proposed model, which includes natural language processing technology was 89.17%. Overall, a 9% improvement was observed.
FMProjector: A Formal Verification Framework for an Operating System Complying with a Standard Interface
http://doi.org/10.5626/JOK.2019.46.8.814
Formal verification techniques facilitate the verification of functional correctness of software. The verification, however, is rarely applicable to large-scale software, such as operating systems, because of the state explosion problem. International standards or certifications, such as IEC-61508 or DO-178, highly recommend formal verification of such software according to the level of safety. The paper introduces a formal verification framework, FMProjector, for operating systems complying with a standard interface. The framework includes horizontal and vertical approaches for systematic analysis of the software based on traceability from the standard interface to the source code. The paper also introduces a case study for the application of FMProjector to Qplus-AIR complying with ARINC-653 which is a standard interface for avionics real-time operating system.
Ship Detection using CNN based on Contrast Fusion Technique in Satellite Images : Accuracy Enhancement
Sunggyun Im, Youngbae Jeon, Junghwan Hwang, Jiwon Yoon
http://doi.org/10.5626/JOK.2019.46.8.823
The satellite has various missions such as ground/marine observation, communication, broadcasting, etc. Satellite photographs provide information for the maintenance of marine security and traffic control for ship detection. Since satellite photos are taken all over the earth, the memory storage is not sufficient to hold such data with each data being of a high resolution and requiring automatic ship detection using the computer. The existing literature on ship detection employed several deep learning models. However, the problem of processing speed due to the characteristics of satellite photographs leads to the necessity of using a CNN(Convolution Neural Network) model that has a comparably high processing speed. On the contrary, it is difficult to improve the accuracy and performance mostly due to factors such as marina, lighthouses and waves. Therefore, in this paper, we propose a model that improves the accuracy and performance by combining image contrast enhancement with the existing CNN. In addition, we have employed the overlap and rotation functions to increase the amount of data required for ship classification in the learning stage and implement automation detection technology considering window sliding to reduce detection speed in real satellite photographs. Also, the identified ship data has been used as learning data to improve accuracy for the model that can be used in the real industry.
Design of a Low-power Rendering Interface for Mobile Augmented Reality Headsets
Jaewon Choi, Hyeonjung Park, JeongGil Ko
http://doi.org/10.5626/JOK.2019.46.8.834
Augmented Reality (AR) headsets such as Microsoft HoloLens are demonstrated as the subsequent emergence of AR applications. GPU is one of the major power consumers in the mobile devices according to prior research works. Especially in AR headset environment, the demand for GPU power is expected to be of more importance because of its always-on nature with 3D graphics rendering burden. In this work, we present Low-power Rendering Interface for mobile augmented reality headsets. The proposed system collects all the graphics call s issued from the application layer and optimizes and applies the graphics calls with respect to user quality preservation by considering user head direction information. All the procedures proceed for 90 us and our system reduces ～27% of power consumption in the best cases.
An Efficient Continuous Subgraph Matching Scheme Considering Data Reuse
Dojin Choi, Kyoungsoo Bok, Jaesoo Yoo
http://doi.org/10.5626/JOK.2019.46.8.842
With an increase in the utilization of graph streams in various applications, a continuous subgraph matching scheme is required to search the subgraphs that undergo changes in real time. In this paper, we propose an efficient continuous subgraph matching scheme that reuses indexing and performs distributed processing in graph stream environments. In order to perform distributed processing, we propose a query decomposition method based on the degree and subsequently manage the decomposed subqueries as an index. The proposed scheme reuses indexing information to reduce the load on the index caused by the environment in which multiple queries are entered. We also conduct query allocation through a cost model that calculates the indexing load of each server. For efficient performance of distributed processing in stream environments, the proposed scheme was implemented in Storm. Various performance evaluations were conducted to demonstrate the superiority of the proposed scheme.
Research and Development of Wireless Protocol Automatic Analyzer
Woorim Bang, Youngbae Jeon, Shinwoo Shim, Kwangsoo Kim, Ji Won Yoon
http://doi.org/10.5626/JOK.2019.46.8.852
Automatic Protocol Reverse Engineering (APRE) defines automatic analysis of the format, semantics, and parameters of an unknown protocol. APRE can be used to detect malware that is distributed on the network, or for security and suitability verification of protocols that have been defined own their own. Conventional APRE studies have been conducted mostly on text-based protocols and wired protocols. As the number of wireless devices increases, there is an increasing need for a protocol analyzer for wireless protocols. Therefore, in this paper, research and development of the protocol automatic analyzer were performed by considering the characteristics of the wireless protocols. For the analysis of the wireless protocol, this study analyzed the messages in binary units. We propose a method to calculate the message distance by assigning a weight according to the packet acquisition time interval to perform clustering among similar messages. As a result of collecting and analyzing the messages according to the IEEE 802.11 protocol using the proposed method, we could correctly classify 95.1% message types among 800messages, and the degree of conciseness was 3.6. By using one of the existing APRE tools, Netzob, 92.1% precision was obtained with the conciseness of 3.5. Consequently, the proposed method showed better performance than Netzob.
An Enhanced Connection Control for Data Connectivity in 3GPP 5G System
http://doi.org/10.5626/JOK.2019.46.8.861
This paper proposes an enhanced connection control for data connectivity in 3GPP 5G System. The 5G System has been developed recently and Rel-15 5G System phase 1 has been released in 3GPP standard. On the other hand, congestion control schemes have been studied on Rel-10 LTE systems based on backoff mechanism for mobility management and session management to control heavy traffic and avoid network overload situation. Also, PS Data off feature has been introduced to prevent activation of all the IP data packets by the user. In this paper, we briefly present the Rel-15 5G System with a connection control and PS-Data off from the perspective and describe the proposed enhanced connection control, called ECC(Enhanced Connection Control). Simulation results for the existing legacy mechanism in 3GPP and the proposed ECC are provided and in future work, improvements in 3GPP 5G System phase 2 standard are considered.
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