• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.1
• /
• pp.77-92
• /
• 2020

When automatic speech recognition (ASR) is provided as a cloud service, it is easy to collect voice and application domain data from users. Harnessing these data will facilitate the provision of more personalized services. In this paper, we demonstrate our transfer learning-based knowledge service that built with the user-generated data collected through our novel system that deliveries personalized ASR service. First, we discuss the motivation, challenges, and prospects of building up such a knowledge-based service-oriented system. Second, we present a Quadruple Transfer Learning (QTL) method that can learn a classification model from a source domain and transfer it to a target domain. Third, we provide an overview architecture of our novel system that collects voice data from mobile users, labels the data via crowdsourcing, utilises these collected user-generated data to train different machine learning models, and delivers the personalised real-time cloud services. Finally, we use the E-Book data collected from our system to train classification models and apply them in the smart TV domain, and the experimental results show that our QTL method is effective in two classification tasks, which confirms that the knowledge transfer provides a value-added service for the upper-layer mobile applications in different domains.

• Journal of Korea Spatial Information System Society
• /
• v.6 no.2 s.12
• /
• pp.51-62
• /
• 2004

Recently, the location-based service for moving customers is becoming one of the most important service in mobile communication area. For moving object applications, there are lots of update operations and such update loads are concentrated on some particular area unevenly. The primary processing of LBS application is spatio-temporal range queries. To improve the throughput of spatio-temporal range queries, the time of disk I/O in query processing should be reduced. In this paper, we adopt non-uniform two-level grid index structures of GALIS architecture,which are designed to minimize update operations. We propose query scheduling technique using spatial relationship and time relationship and a combined spatio-temporal query processing method using time zone concepts to improve the throughput of query processing. Some experimental results are shown for range queries with different query range to show the performance tradeoffs of the proposed methods.

• Journal of Technology Innovation
• /
• v.22 no.1
• /
• pp.1-22
• /
• 2014

R&D investment also has a gestation period similar to other investments in economics. The gestation period originates from time lag effect of input and output. Thus it is necessary to consider time lag effects when analyzing the relationship between firms' R&D investment and R&D performance. The main objective of this research is to estimate the length of time lag effect of R&D investment. The Almon distribution lag model was applied to estimate the time lag effect. The firm level panel dataset was established from 2002 to 2009. The net value of R&D investment and the number of patent applications were used to measure R&D input and output, respectively. This method found the estimated time lag to be 1~2 years across all datasets. The same analyses were applied to chemical, metal, electronic, exact science, and machinery industries' data. And we found there were differences among sectors in regard to the time lag effect.

• Journal of Korea Multimedia Society
• /
• v.13 no.4
• /
• pp.575-581
• /
• 2010

With widespread use of the Internet and improvements in streaming media and compression technology, digital music, video, and image can be distributed instantaneously across the Internet to end-users. However, most conventional Digital Right Management are often not secure and not fast enough to process the vast amount of data generated by the multimedia applications to meet the real-time constraints. The SVC offers temporal, spatial, and SNR scalability to varying network bandwidth and different application needs. Meanwhile, for many multimedia services, security is an important component to restrict unauthorized content access and distribution. This suggests the need for new cryptography system implementations that can operate at SVC. In this paper, we propose a new scrambling encryption for reserving the characteristic of scalability in MPEG4-SVC. In the base layer, the proposed algorithm is applied and performed the selective scambling. And it encrypts various MVS and intra-mode scrambling in the enhancement layer. In the decryption, it decrypts each encrypted layers by using another encrypted keys. Throughout the experimental results, the proposed algorithms have low complexity in encryption and the robustness of communication errors.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.24 no.6B
• /
• pp.1098-1107
• /
• 1999

In multimedia applications, a source usually generates multiple streams. By heterogeneous multimedia multicast, we mean a recipient can receive some of them, not necessarily all of them. A recipient bids for what it wants to receive and the source gains the same amount when a connection is established. The problem of distributing streams for heterogeneous multicast to maximize the source's gain, can be solved using a 0-1 integer programming, hewn as NP-complete. In this paper, we propose efficient stream distribution algorithms in two different types of multicast models. The first restricted model assumes that the capacity for a link in the multicast tree is grater than or equal to the capacities of its descendant links. In the second unrestricted model, we drop out the restriction in the restricted model. Proposed algorithms have better time and space complexities compared with any existing one. In addition, distributed implementations are straightforward, which is very useful for large networks.

• Proceedings of the Korean Institute of IIIuminating and Electrical Installation Engineers Conference
• /
• 2004.11a
• /
• pp.371-374
• /
• 2004

Induction motors are a critical component of many industrial processes and are frequently integrated in commercially available equipment. Safety, reliability, efficiency, and performance are some of the major concerns of induction motor applications. Preventive maintenance of induction motors has been a topic great interest to industry because of their wide range application of industry. Since the use of mechanical sensors, such as vibration probes, strain gauges, and accelerometers is often impractical, the motor current signature analysis (MACA) techniques have gained murk popularity as diagnostic tool. Fault tolerant control (FTC) strives to make the system stable and retain acceptable performance under the system faults. All present FTC method can be classified into two groups. The first group is based on fault detection and diagnostics (FDD). The second group is independent of FDD and includes methods such as integrity control, reliable stabilization and simultaneous stabilization. This paper presents the fundamental FDD-based FTC methods, which are capable of on-line detection and diagnose of the induction motors. Therefore, our group has developed the embedded distributed fault tolerant and fault diagnosis system for industrial motor. This paper presents its architecture. These mechanisms are based on two 32-bit DSPs and each TMS320F2407 DSP module is checking stator current, voltage, temperatures, vibration and speed of the motor. The DSPs share information from each sensor or DSP through DPRAM with hardware implemented semaphore. And it communicates the motor status through field bus (CAN, RS485). From the designed system, we get primitive sensors data for the case of normal condition and two abnormal conditions of 3 phase induction motor control system is implemented. This paper is the first step to drive multi-motors with serial communication which can satisfy the real time operation using CAN protocol.

• The KIPS Transactions:PartA
• /
• v.10A no.6
• /
• pp.651-656
• /
• 2003

Emerging technologies, such as multimedia and mobile computing, require that middleware platforms can support high adaptability in order to provide more flexible services in heterogeneous environments. To support high quality of service in distributed wired/wireless environments, it will be necessary for applications and middleware to be aware of changes in users requirements as well as environmental conditions, also to be able to adapt their behaviour as such changes. According to the needs of adaptability supporting in middleware, we structure a component-based, mobile agents framework in base level and meta level by using reflection. We propose concepts of meta agents and meta-service agents that are able to reflect changes made by mobile agents to the system. At the meta level of our framework, meta agents monitor execution of mobile agents and it is possible to provide dynamic adaptation of users requirements, deployment of application services and service customization with meta agents and meta-service agents.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.1
• /
• pp.9-32
• /
• 1996

In this paper, we describe a multimedia data processing model that supports wide variety of applications based on multimedia production model. This model supports network-transparent access to stored multimedia data, real-time mult media input devices, and multimedia processing. The model addresses real-time dato a switching and delivery, as well as acquisition, processing, and output. Most translation, compression, and synchronization services are integrated. This model consists of three layers: (1) stream, (2) multimedia presentation, (3) hyperpresentaion. This paper describes the data abstractions associated with each layer. These data abstractions provide a framework for defining the services provided by each layer, and describe the object-oriented mechanisms that provide those services. A sample scenario is presentated to illustrate the use of this model. A server-client architecture and implementation issues, and future directions are also discussed. sed.

• Journal of KIISE:Computing Practices and Letters
• /
• v.10 no.1
• /
• pp.49-56
• /
• 2004

Load imbalance is a serious impediment to achieving good performance in parallel processing. Global load balancing schemes cannot adequately manage to balance parallel tasks generated from a single application. Dynamic loop scheduling methods are known to be useful in balancing parallel loops on shared-memory multiprocessor machines. However, their centralized nature causes a bottleneck for the relatively small number of processors in a network of workstations because of order-of-magniture differences in communication overheads. Moreover, improvements of basis loops scheduling methods have not effectively dealt with irregularly distributed workloads in parallel loops, which commonly occur in applications for a network of workstation. In this paper, we present a new reconfigurable and decentralized balancing method for parallel loops on a network of workstations. Since our method supplements performance balancing with those tranditional load balancing methods, it minimizes the overall execution time.

• Journal of KIISE:Software and Applications
• /
• v.31 no.8
• /
• pp.1043-1053
• /
• 2004

The performance of a component can significantly influence the overall performance of the system being developed in the component based software development. Existing performance measurement tools for components are often built specific to the component model or middleware. One disadvantage of this approach is that when the system is moved to a new platform during the life-cycle of the system, the measurement tools also need to be adapted. A similar problem is observed when components built for different platforms are integrated with a new integration technology like Web Services. This paper presents a new model for measuring component performances, which is independent of particular component models or middleware. This method presents a interoperability for each performance data by model transformation process from middleware independent performance measuring models to middleware dependent performance measuring models. Our model allows objective performance data to be obtained that can be used to compare performances of components based on different component models or middleware platforms.