• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권1호
• /
• 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 Welding and Joining
• /
• 제20권5호
• /
• pp.127-133
• /
• 2002

In this study, metal transfer characteristics in pulsed current metal inert gas (MIG) welding of aluminum was investigated. Based on the metal transfer characteristics from direct current electrode negative MIG welding, the one drop per one pulse(ODOP) condition was predicted and compared with experimental data. The results indicated that experimental pulse range for the ODOP condition is wider than that predicted from the DCEP MIG welding data. In addition, more stable metal trnasfer behavior was obtained at the higher end of the ODOP condition.

• Investigative Magnetic Resonance Imaging
• /
• 제25권1호
• /
• pp.10-22
• /
• 2021

Purpose: To overcome the difficulty in building a large data set with a high-quality in medical imaging, a concept of 'blended-transfer learning' (BTL) using a combination of both source data and target data is proposed for the target task. Materials and Methods: Source and target tasks were defined as training of the source and target networks to reconstruct cardiac CINE images from undersampled data, respectively. In transfer learning (TL), the entire neural network (NN) or some parts of the NN after conducting a source task using an open data set was adopted in the target network as the initial network to improve the learning speed and the performance of the target task. Using BTL, an NN effectively learned the target data while preserving knowledge from the source data to the maximum extent possible. The ratio of the source data to the target data was reduced stepwise from 1 in the initial stage to 0 in the final stage. Results: NN that performed BTL showed an improved performance compared to those that performed TL or standalone learning (SL). Generalization of NN was also better achieved. The learning curve was evaluated using normalized mean square error (NMSE) of reconstructed images for both target data and source data. BTL reduced the learning time by 1.25 to 100 times and provided better image quality. Its NMSE was 3% to 8% lower than with SL. Conclusion: The NN that performed the proposed BTL showed the best performance in terms of learning speed and learning curve. It also showed the highest reconstructed-image quality with the lowest NMSE for the test data set. Thus, BTL is an effective way of learning for NNs in the medical-imaging domain where both quality and quantity of data are always limited.

• 설비공학논문집
• /
• 제15권12호
• /
• pp.1007-1017
• /
• 2003

An experimental study on the air-side pressure drop and heat transfer coefficient of slit fin-tube heat exchanger has been carried out. The data reduction methodology for air-side heat transfer coefficients in the literature is not based on a consistent approach. This paper focuses on new method of data reduction to obtain the air-side performance of fin-tube heat exchanger using R22 and recommends standard procedures for dry and wet surface heat transfer estimation in fin-tube heat exchanger having refrigerant on the tube-side. Results are presented as plots of friction f-factor and Colburn j -factor against Reynolds number based on the fin collar outside diameter and compared with previous studies. The data covers a range of refrigerant mass fluxes of 150∼250 kg/$m^2$s with air flows at velocity ranges from 0.3 m/s to 0.8 m/s.

• 설비공학논문집
• /
• 제16권5호
• /
• pp.490-498
• /
• 2004

This paper measures the forced convective heat transfer from heated cylinder to air flow in a rectangular duct at Re$_{D}$ =2,337, 4,589, 6,621 and 7,944 through experiments. And the heat transfer is computed by 3-D numerical computation in which various turbulent models are applied. It is shown through the comparison of experimental and computed data that numerical computation with standard k-$\varepsilon$ model predicts the experimental data most accurately. Furthermore, the correlation from the computed heat transfer is almost similar to that from the experiment when Re$_{D}$ is greater than 4,589. In addition, the correlation of McAdams is the closest to that from experimental data among various correlations from literature in the range of Reynolds number.ber.

• 설비공학논문집
• /
• 제15권3호
• /
• pp.220-229
• /
• 2003

An experimental study on the air-side pressure drop and heat transfer coefficient of slit fin-tube heat exchanger has been carried out. The data reduction methodology for air-side heat transfer coefficients in the literature is not based on a consistent approach. This paper focuses on new method of data reduction to obtain the air-side performance of fin-tube heat exchanger using R22 and recommends standard procedures for dry surface heat transfer estimation in fin-tube heat exchanger having refrigerant on the tube-side. Results are presented as plots of friction f-factor and Colburn j -factor against Reynolds number based on the fin collar outside diameter and compared with previous studies. The data covers a range of refrigerant mass fluxes of 150~250 kg/$m^2$s with air flows at velocity ranges from 0.6 m/s to 1.6 m/s.

• 전자공학회논문지B
• /
• 제33B권9호
• /
• pp.70-79
• /
• 1996

Block data transfer poses a serious problem is a pipelined bus where each data transfer step is pipelined. In this paper, we describe the design and implementation of a variable data block transfer protocol for a pipelined bus of a shared-memory multiprocessor. The proposed method maintains compatibility with the existing protocol for the pipelined bus and ensures fairness and effectiveness by preventing starvation. We present flow charts of requester and responder during a block transfer in the pipelined bus that uses the proposed protocol. The proposed protocol was implemented for the TICOM-III HiPi+Bus.

• International Journal of Air-Conditioning and Refrigeration
• /
• 제9권4호
• /
• pp.27-36
• /
• 2001

A new two-phase non-boiling convective heat transfer correlation for turbulent flow $(Re_{SL}>4000)$ in vertical tubes with different fluid flow patterns and fluid combinations was developed using experimental data available from the literature. The correlation presented herein originates from a careful analysis of the major non-dimensional parameters affecting two-phase heat transfer. This model takes into account the appropriate contributions of both the liquid and gas phases using the respective cross-sectional areas occupied by the two phases. A total of 255 data points from three available studies (which included the four sets of data) were used to determine the curve-fitted constants in the improved correlation. The performance of the new correlation was compared with two-phase correlations from the literature, which were developed for specific fluid combinations.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 추계학술대회
• /
• pp.354-356
• /
• 2021

대용량 데이터의 공유 및 신속한 전송을 요구하는 과학응용 분야에서는 전송 성능을 높이기 위해 Globus/GridFTP와 같은 전송 어플리케이션에서의 병행성, 병렬성, 파이프라이닝 기법 등을 통한 성능 향상을 추구해 오고 있다. 본 논문에서는 비슷한 맥락에서 전송 프로토콜 계층에서 다중 경로 지원할 수 있는 Mptcp를 도입할 경우 기대할 수 있는 인터페이스 수의 증가 및 전송 스트림들의 병렬성 확장에 따른 전송성능 향상을 에뮬레이션 환경에서 실험하고 결과를 제시한다.

• Journal of Advanced Marine Engineering and Technology
• /
• 제28권6호
• /
• pp.1026-1036
• /
• 2004

In this study, single-phase heat transfer experiments were conducted with Oblong Shell and Plate heat exchanger using water. An experimental water loop has been developed to measure the single-phase heat transfer coefficient and pressure drop in a vertical Oblong Shell and Plate heat exchanger. Downflow of hot water in one channel receives heat from the cold water upflow of water in the other channel. Similar to the case of a plate heat exchanger, even at a very low Reynolds number, the flow in the Oblong Shell and Plate heat exchanger remains turbulent. The present data show that the heat transfer coefficient and pressure drop increase with the Reynolds number. Based on the present data, empirical correlations of the heat transfer coefficient and pressure drop in terms of Nusselt number and friction factor were proposed. Also, performance prediction analyses for Oblong Shell and Plate heat exchanger were executed and compared with experiments. $\varepsilon$-NTU method was used in this prediction program. Independent variables are flow rates and inlet temperatures. Compared with experimental data, the accuracy of the program is within the error bounds of $\pm$5% in the heat transfer rate.