Vol. 47, No. 4,
Apr. 2020
Digital Library
Branchpoint Prediction Using Self-Attention Based Deep Neural Networks
http://doi.org/10.5626/JOK.2020.47.4.343
Splicing is a ribonucleic acid (RNA) process of creating a messenger RNA (mRNA) translated into proteins. Branchpoints are sequence elements of RNAs essential in splicing. This paper proposes a novel method for branchpoint prediction. Identification of branchpoints involves several challenges. Branchpoint sites are known to depend on several sequence patterns, called motifs. Also, a branchpoint distribution is highly biased, imposing a class-imbalanced problem. Existing approaches are limited in that they either rely on handcrafted sequential features or ignore the class imbalance. To address those difficulties, the proposed method incorporates 1) Attention mechanisms to learn sequence-positional long-term dependencies, and 2) Regularization with triplet loss to alleviate the class imbalance. Our method is comparable to the state-of-the-art performance while providing rich interpretability on its decisions.
Scenario-Based Multi-core Multi-tasking of Engine Control Unit Real-Time Object-Oriented Model
http://doi.org/10.5626/JOK.2020.47.4.352
This paper proposes a method for the efficient execution of real-time object-oriented models in multi-core systems using a vehicle Engine Control Unit (ECU) system as a case study. All current commercial and open source real-time object-oriented modeling tools use objects as basic units of mapping to tasks and cores. In contrast, we propose a method of using scenarios instead of objects as the basic units of mapping. To demonstrate the efficiency of the proposed method, we used an ECU system model and eTrice, which is the only free open source among real-time object-oriented modeling tools. Specifically, we implemented an ECU real-time object-oriented model in eTrice, and extended the implementation of eTrice Linux run-time systems and code generators to support scenario-based architecture and multi-core synchronization mechanisms. Our results demonstrate that the proposed scenario-based multi-core multi-tasking improves both theoretical and experimental response spare times (slacks) compared with current and previous work.
RNN-based Body Posture Estimation Method for 3D UI in Virtual Reality
Wonjun Seong, Byungmoon Kim, BoYu Gao, Jini Kwon, HyungSeok Kim
http://doi.org/10.5626/JOK.2020.47.4.362
Providing intuitive and natural interface is crucial in virtual reality applications. There have been many studies on proper 3D interfaces that require the accurate tracking of body posture. Based on previous studies, we propose an effective body posture estimation method for 3D interfaces. In this study, we focused on the body-fixed UI, especially movements of the upper body. To track body posture, it is necessary to apply additional sensors such as RGB/RGB-D camera, magnetic/optical trackers, etc. The goal of this study was to track the body posture with conventional virtual reality devices only. We applied conventional HMD with head tracker and hand-held controllers. With these three trackers which are not directly attached to the body, an RNN-based method is proposed to effectively estimate the upper body pose. Experiments shows that the proposed method could track the upper body position and orientation within 5% of the error rate, compared with the explicit tracker data. The proposed method could be applied to design body-fixed UI or body-related interfaces without additional devices, which would enhance the accessibility of virtual reality applications.
A Reference Architecture for Machine Learning-Based Autonomous Systems
http://doi.org/10.5626/JOK.2020.47.4.368
Autonomous computing is one of the essential factors for realizing the fourth industrial revolution and a future technology that provides capabilities of autonomous recognition, autonomous judgement, autonomous planning, and autonomous management with automatic systems. With the advent of various sensors and IoT devices, a rich set of context data can be acquired from the environment, and autonomous system technologies with human-machine interface (HMI) enabling the realization of an eco-system wherein a system itself can maintain its best quality by using the acquired context data. However, because of the highly complicated functional and non-functional requirements for realizing autonomous systems, developing such systems becomes more difficult and development productivity becomes much lower. In the paper, we present a reference architecture which can be commonly applied to autonomous systems. The proposed reference architecture includes architecture design, core components, main algorithm, and so on. The reference architecture forms a structural basis of the target system and can guarantee the overall quality and improve development efficiency by reusing the core structure of the reference architecture. Additionally, we apply the reference architecture to two autonomous systems and verify the applicability and practicability of the reference architecture.
Analysis of Utilization Methods of the Statistical Model Checking Results for Localizing Faults on System of Systems
Sangwon Hyun, Yong-jun Shin, Doo-Hwan Bae
http://doi.org/10.5626/JOK.2020.47.4.380
System of Systems (SoS) is a large and complex system comprising independent constituent systems. Statistical Model Checking (SMC) techniques can be used to verify if the SoS achieves its goals or not. However, even if the SoS goal failure is detected using the SMC, finding a root cause of the SoS failure requires more cost than that of a system. One of the candidate solutions for reducing the debugging cost is to apply fault localization techniques on the SoS. However, existing fault localization techniques are designed to utilize testing results of a system. Thus, a method to utilize SMC results is needed to apply existing fault localization techniques to the SoS. In this study, we suggest six utilization methods of SMC results for SoS fault localization, and compare the performance of them on the emergency-response SoS. We found that the method based on the expectation value showed the best performance in the experiment.
Data-Driven Computer-Aided Diagnosis of Ventricular Fibrillation Based on Ensemble Empirical Mode Decomposition of ECG
http://doi.org/10.5626/JOK.2020.47.4.387
In this paper, we propose a novel computer-aided diagnosis method to detect VF(ventricular fibrillation), one of the hazardous cardiac symptoms of arrhythmia by applying the EEMD(Ensemble Empirical Mode Decomposition) to the ECG signals. Using the EEMD to the ECG signals, it is shown that VF in the EMD region has a higher correlation with the IMFs (intrinsic mode functions) than the NSR (normal sinus rhythm) and other types of arrhythmia. To quantify this characteristic, we calculate the angle between the ECG signal and the specific IMFs, and classify the pathology by differentiating the angles. To verify the effectiveness of the proposed algorithm, we measured the accuracy of diagnosis using arrhythmia data from the PhysioNet database and confirm capacity of the proposed method.
Morpheme-based Korean Word Vector Generation Considering the Subword and Part-Of-Speech Information
http://doi.org/10.5626/JOK.2020.47.4.395
Word vectors enable finding the relationship between words by vector computation. They are also widely used as pre-trained data for high-level neural network programs. Various modified models from English models have been proposed for the generation of Korean word vectors, with various segmentation units such as Eojeol(word phrase), morpheme, syllable and Jaso. In this study, we propose Korean word vector generation methods that segment Eojeol into morphemes and convert them into subwords comprising either syllable or Jaso. We also propose methods using Part-Of-Speech tags provided in the pre-processing to reflect semantic and syntactic information regarding the morphemes. Intrinsic and extrinsic experiments showed that the method using morpheme segments with Jaso subwords and additional Part-Of-Speech tags showed better performance than others under the condition that the target data are normal text and not as grammatically incorrect.
Ontology and CNN-based Inference of the Threat Relationship Between UAVs and Surrounding Objects
MyungJoong Jeon, MinHo Lee, HyunKyu Park, YoungTack Park, Hyung-Sik Yoon, Yun-Geun Kim
http://doi.org/10.5626/JOK.2020.47.4.404
The technology that identifies the relationship between surrounding objects and recognizes the situation is considered as an important and necessary technology in various areas. Numerous methodologies are being studied for this purpose. Most of the studies have solved the problem by building the domain knowledge into ontology for reasoning of situation awareness. However, based on the existing approach; it is difficult to deal with new situations in the absence of domain experts due to the dependency of experts on relevant domain knowledge. In addition, it is difficult to build the knowledge to infer situations that experts have not considered. Therefore, this study proposes a model for using ontology and CNN for reasoning of the relationships between UAVs and surrounding objects to solve the existing problems. Based on the assumption that the accuracy of ontology reasoning is insufficient, first, the reasoning was performed using the information from the detected surrounding objects. Later, the results of ontology reasoning are revised by CNN inference. Due to the limitations of actual data acquisition, data generator was built to generate data similar to real data. For evaluation of this study, two models of relationships between two objects were built and evaluated; both the models showed over 90% accuracy.
Passage Re-ranking Method Based on Sentence Similarity Through Multitask Learning
Youngjin Jang, Hyeon-gu Lee, Jihyun Wang, Chunghee Lee, Harksoo Kim
http://doi.org/10.5626/JOK.2020.47.4.416
The machine reading comprehension(MRC) system is a question answering system in which a computer understands a given passage and respond questions. Recently, with the development of the deep neural network, research on the machine reading system has been actively conducted, and the open domain machine reading system that identifies the correct answer from the results of the information retrieval(IR) model rather than the given passage is in progress. However, if the IR model fails to identify a passage comprising the correct answer, the MRC system cannot respond to the question. That is, the performance of the open domain MRC system depends on the performance of the IR model. Thus, for an open domain MRC system to record high performance, a high performance IR model must be preceded. The previous IR model has been studied through query expansion and reranking. In this paper, we propose a re-ranking method using deep neural networks. The proposed model re-ranks the retrieval results (passages) through multi-task learning-based sentence similarity, and improves the performance by approximately 8% compared to the performance of the existing IR model with experimental results of 58,980 pairs of MRC data.
Cascading Behavior and Information Diffusion in Overlapping Clusters
Woojung Lee, Joyce Jiyoung Whang
http://doi.org/10.5626/JOK.2020.47.4.422
Information diffusion models formulate and explain cascading behavior in networks where a small set of initial adopters is assumed to acquire new information and the new information is propagated to the other nodes in the network. Most existing information diffusion models assume that a node in a network belongs to only one cluster, and based on this assumption, it has been shown that clusters are obstacles to cascades. However, in many real-world networks, a node can belong to multiple clusters, i.e., clusters can overlap. In this paper, we study cascading behavior in a network when clusters overlap. We show that clusters are not obstacles to cascades if the initial adopters are placed in the overlapped region between the clusters or if we allow compatibility. We verify our theorems and models on four real-world datasets.
Design and Implementation of Indoor Positioning System Using Particle Filter Based on Wireless Signal Intensity
http://doi.org/10.5626/JOK.2020.47.4.433
This paper proposes an Indoor Positioning System to track a user’s position indoors by using beacons’ wireless signal intensity. To overcome the non-linearity of an existing indoor positioning scheme using wireless signal intensity, a particle filter is used for a positioning algorithm, so the noise of the wireless signal intensity is not directly reflected on the positioning result. In the observation phase of the particle filter, the distance from a user’s smartphone is estimated based on the wireless signal intensity, and the similarity of each particle with an estimated ground truth is calculated through the predicted distance value. Also, our proposed positioning scheme uses the random walk technique (the Monte Carlo method) to calculate a position estimation value. Additionally, to solve the well-known local minimum problem of the particle filter, the particles estimated closest to the beacons according to the distance prediction values are given proximity weights, so the particles can quickly locate the user. The positioning error on the walking path is also corrected by considering the indoor map.
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