• Journal of Sensor Science and Technology
• /
• v.30 no.5
• /
• pp.290-294
• /
• 2021

Experts have designed popular and successful model architectures, which, however, were not the optimal option for different scenarios. Despite the remarkable performances achieved by deep neural networks, manually designed networks for classification tasks are the backbone of object detection. One major challenge is the ImageNet pre-training of the search space representation; moreover, the searched network incurs huge computational cost. Therefore, to overcome the obstacle of the pre-training process, we introduce a network adaptation technique using a pre-trained backbone model tested on ImageNet. The adaptation method can efficiently adapt the manually designed network on ImageNet to the new object-detection task. Neural architecture search (NAS) is adopted to adapt the architecture of the network. The adaptation is conducted on the MobileNetV2 network. The proposed NAS is tested using SSDLite detector. The results demonstrate increased performance compared to existing network architecture in terms of search cost, total number of adder arithmetics (Madds), and mean Average Precision(mAP). The total computational cost of the proposed NAS is much less than that of the State Of The Art (SOTA) NAS method.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.2E
• /
• pp.66-71
• /
• 2009

In practical applications of speech recognition, one of the fundamental questions might be on the number of training data that should be provided for a specific task. Though plenty of training data would undoubtedly enhance the system performance, we are then faced with the problem of heavy cost. Therefore, it is of crucial importance to determine the least number of training data that will afford a certain level of accuracy. For this purpose, we investigate the effect of the number of training data on the speaker-independent speech recognition of isolated words by using FVQ/HMM. The result showed that the error rate is roughly inversely proportional to the number of training data and grows linearly with the vocabulary size.

• Korean Journal of Computational Design and Engineering
• /
• v.16 no.5
• /
• pp.370-379
• /
• 2011

The increasing complexity of complex buildings, such as high-rise buildings and underground subway stations, presents new challenges to firefighters. In a fire in complex buildings, the importance of the collaboration between firefighters is clear. The increased demand on firefighter training for such environment is now evident. Due to cost, time, and safety issues, it is impossible to experience a real fire in such environments for training. In addition, the use of real fire for training does not enable repeatable training and the evaluation of the training is difficult. We developed a team-based firefighter training simulator for complex buildings using the virtual reality technology. It provides the training and evaluation of firefighting and mission-based team training. To model real fire phenomena in virtual space, a numerical analysis method based on fire dynamics is used. To achieve an immersive virtual environment, an augmented reality technique for the compensation of real world image and a haptic technique for heat experience are adopted. The developed training simulator can help the firefighter to respond to large and complex firefighting scenarios, while maintaining the safety of the trainees.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.17 no.31
• /
• pp.11-20
• /
• 1994

In these days, as human resource development is emphasized in an enterprise, the importance of evaluation of the education and training which is the means to practice is gradually increased. Because people wants to know the effects compared to the cost of the education and training in several years. Therefore, in this paper, I testified to a contribution about the strategic purpose of the functions of the education and training and I represented several strategies to practice it Through these strategies. the department of the education and training is getting know that how much does the education and training affect a prodectivity and a profit in offering service. And I also represented that it can be applied soft system analysis in developing concept model of ways of doing things. This analysis explains detaily the job performance skills which needs in each constituent element of system. The education and training can develop the ways of improving the job performance skills. The job performance skills affect a productivity and a profit. Finally, increasing of a profit due to the education and training represents the contribution of the strategic purpose and emphasize the development of the program for it. The education and training have to be considered the development of the marketing plan. If the prefects of the department of the education and training help their strategic purpose achievement, other departments can be given the services offering the education and training. The education and training measures the effects, embodies them and needs to be more sensitive making datas of the success cases.

• Archives of Plastic Surgery
• /
• v.41 no.3
• /
• pp.218-224
• /
• 2014

Plastic surgery training worldwide has seen a thorough restructuring over the past decade, with the introduction of formal training curricula and work-based assessment tools. Part of this process has been the introduction of revalidation and a greater use of simulation in training delivery. Simulation is an increasingly important tool for educators because it provides a way to reduce risks to both trainees and patients, whilst facilitating improved technical proficiency. Current microsurgery training interventions are often predicated on theories of skill acquisition and development that follow a 'practice makes perfect' model. Given the changing landscape of surgical training and advances in educational theories related to skill development, research is needed to assess the potential benefits of alternative models, particularly cross-training, a model now widely used in non-medical areas with significant benefits. Furthermore, with the proliferation of microsurgery training interventions and therefore diversity in length, cost, content and models used, appropriate standardisation will be an important factor to ensure that courses deliver consistent and effective training that achieves appropriate levels of competency. Key research requirements should be gathered and used in directing further research in these areas to achieve on-going improvement of microsurgery training.

• Asian-Australasian Journal of Animal Sciences
• /
• v.6 no.1
• /
• pp.91-97
• /
• 1993

This experiment was carried out to examine the effects of urea-treated rice straw feeding on the growth performance of crossbred Holstein heifers under different feeding conditions. In the first experiment, the animals were given diets having 2 levels of TDN and CP and 3 levels of crude fiber (22, 30 and 36%) which were formulated with urea-treated rice straw and concentrates. Daily weight gain of heifers was not different between 22% and 30% CF diets, but the reduction of TDN or CP level to 90% of the requirements decreased the weight gain. Fiber content of 36% also reduced the body weight gain. The reduction of TDN significantly reduced DM intake and increased feed conversion ratio. Feed cost per kg weight gain was significantly increased by an increase in CF to 36%. In the second experiment, separate feeding and total mixing feeding were compared. There were no significant differences between the two feeding systems in body weight gain although the possibility of superiority in SF to TMF remained. DM intake was not affected by the feeding system, but 30% CF diet gave higher DM intake. Feed cost per kg weight gain was lower in the 30% CF diet.

• Journal of Multimedia Information System
• /
• v.7 no.4
• /
• pp.277-280
• /
• 2020

Extended reality-assisted training offers repeatable learning opportunities at a low cost. This paper proposes a smart glass application for training core nursing skills to nursing students, who often need to memorize and practice training sequences. The proposed smart glass application interactively presents a series of instructions to help students remember and perform two core nursing skills in the correct order. We conducted a usability test on 30 undergraduate nursing students in their third year using the smart glass application. Our initial findings show that many students have positively evaluated the possibility of using smart glasses for training, but have also encountered several challenges with the smart glass application's user interface, which takes time to adapt.

• Journal of Advanced Marine Engineering and Technology
• /
• v.40 no.4
• /
• pp.342-347
• /
• 2016

This paper dealt with the economic analysis on the application of shipboard LED (Light Emitting Diode) luminaires to replace incandescent and fluorescent lamps, which account for over 80 % of light source on a training or naval vessel. The rates of savings achieved in the power consumption, initial investment, maintenance cost, and fuel cost were analyzed. The break-even points and the $CO_2$ emissions were also calculated. For the training vessel, the initial investment was increased by 3.8 times, while the maintenance cost over five years was reduced by 51 %. When 40 %, 50 %, and 60 % of luminaires were turned on, the calculated break-even points were 11 months, 9 months, and 7 months, respectively. On the other hand, the initial investment was increased by 3.5 times while the maintenance cost over five years was saved by 55 % for the naval vessel. The break-even points were calculated as 15 months, 12 months, and 10 months, respectively. Moreover, the $CO_2$ emissions of the training and the naval vessels were reduced by 69 % and 65 %, respectively. From the results, it was revealed that the maintenance cost can be reduced by more than 50 %, and that the power consumption, fuel cost, and $CO_2$ emissions can be reduced by more than 60 % if LED luminaires are used on two types of vessels.

• International conference on construction engineering and project management
• /
• 2024.07a
• /
• pp.1322-1322
• /
• 2024

Reinforced concrete (RC) , a major contributor to resource depletion and harmful emissions, fuels research on optimizing its design. Optimizing RC structures is challenging due to the mix of discrete and continuous variables, hindering traditional differentiation-based methods. Thus, this study aims to optimize RC structures cost-effectively using deep reinforcement learning. When the Agent selects design variables, Environment checks design criteria based on KDS 14-20 code (South Korea) and calculates reward. The Agent updates its Neural Network with this reward. Target for optimization is a simply supported doubly RC beam, with design variables including cross-section dimensions, sizes and quantities of tension and compression reinforcement, and size of stirrups. We used 200,000 training sets and 336 test sets, each with live load, dead load, beam length variables. To exclude labeled data, multiple training iterations were conducted. In the initial training, the reward was the ratio of maximum possible cost at beam length to the designed structure's cost. Next iterations used the ratio of optimal values by the previous Agent to the current Agent as the reward. Training ended when the difference between optimal values from the previous and current Agent was within 1% for test data. Brute Force Algorithm was applied to the test set to calculate the actual cost-optimal design for validation. Results showed within 10% difference from actual optimal cost, indicating successful deep reinforcement learning application without labeled data. This study benefits the rapid and accurate calculation of optimized designs and construction processes in Building Information Modeling (BIM) applications.

• Industrial Engineering and Management Systems
• /
• v.12 no.4
• /
• pp.368-379
• /
• 2013

Efficiency in work training is a perennial issue due to high-diversity low-volume production, particularly for manufacturers producing office machines which are manually assembled by workers. To reduce the training cost, parts used in training are usually reused; a trainer disassembles a product assembled by a worker in training. This paper proposes a training method that employs disassembly usually performed by a trainer. This method assigns both assembly and disassembly to a worker in training, in contrast to the conventional method. The effectiveness of the proposed method is experimentally discussed in terms of learning assembly motions and work procedure at each learning stage, namely, "undergoing learning," "immediately after learning," and "seven days after learning." The effectiveness of the training method is confirmed. The method improves the stability of work procedure recollection immediately after training. Furthermore, at seven days after training, it improves retention of the assembly motions and work procedure, and also promotes and maintains memory related to product structure.