• IEMEK Journal of Embedded Systems and Applications
• /
• v.15 no.2
• /
• pp.43-49
• /
• 2020

In this paper, we propose a stage-wise knowledge transfer method that uses block-wise retraining to transfer the useful knowledge of a pre-trained residual network (ResNet) in a teacher-student framework (TSF). First, multiple hint information transfer and block-wise supervised retraining of the information was alternatively performed between teacher and student ResNet models. Next, Softened output information-based knowledge transfer was additionally considered in the TSF. The results experimentally showed that the proposed method using multiple hint-based bottom-up knowledge transfer coupled with incremental block-wise retraining provided the improved student ResNet with higher accuracy than existing KD and hint-based knowledge transfer methods considered in this study.

• ETRI Journal
• /
• v.41 no.2
• /
• pp.242-253
• /
• 2019

We devise a layer-wise hint training method to improve the existing hint-based knowledge distillation (KD) training approach, which is employed for knowledge transfer in a teacher-student framework using a residual network (ResNet). To achieve this objective, the proposed method first iteratively trains the student ResNet and incrementally employs hint-based information extracted from the pretrained teacher ResNet containing several hint and guided layers. Next, typical softening factor-based KD training is performed using the previously estimated hint-based information. We compare the recognition accuracy of the proposed approach with that of KD training without hints, hint-based KD training, and ResNet-based layer-wise pretraining using reliable datasets, including CIFAR-10, CIFAR-100, and MNIST. When using the selected multiple hint-based information items and their layer-wise transfer in the proposed method, the trained student ResNet more accurately reflects the pretrained teacher ResNet's rich information than the baseline training methods, for all the benchmark datasets we consider in this study.

• Journal of information and communication convergence engineering
• /
• v.15 no.2
• /
• pp.91-96
• /
• 2017

PIN-entry interfaces have high risks to leak secret values if the malicious attackers perform shoulder-surfing attacks with advanced monitoring and observation devices. To make the PIN-entry secure, many studies have considered invisible radio channels as a secure medium to deliver private information. However, the methods are also vulnerable if the malicious adversaries find a hint of secret values from user's $na{\ddot{i}}ve$ gestures. In this paper, we revisit the state-of-art radio channel based bimodal PIN-entry method and analyze the information leakage from the previous method by exploiting the sight tracking attacks. The proposed sight tracking attack technique significantly reduces the original password complexities by 93.8% after post-processing. To keep the security level strong, we introduce the advanced bimodal PIN-entry technique. The new technique delivers the secret indicator information through a secure radio channel and the smartphone screen only displays the multiple indicator options without corresponding numbers. Afterwards, the users select the target value by following the circular layout. The method completely hides the password and is secure against the advanced shoulder-surfing attacks.

• Journal of Technology Innovation
• /
• v.19 no.1
• /
• pp.177-201
• /
• 2011

This study is trying to estimate the effect of survey of prior art on the technological performance of national R&D program, with the purpose to enhance understanding on the relationship between utilization of patent information and R&D activities. Patent and Technology Trend Research, one of the survey of prior art which gives the information about existing technology and patent trend to the project team has been carried out since 2005. In this paper, effects which Patent and Technology Trend Research has on the technological performance of national R&D projects are estimated by using multiple regression model considering input factors, characteristics of an agent and supplier of money. The result is that Patent and Technology Trend Research has positive and significant effect on the grant and application of domestic and foreign patent. This result can give an hint that the utilization of patent information make the R&D process efficient and effective.

• Journal of KIISE:Information Networking
• /
• v.27 no.1
• /
• pp.88-97
• /
• 2000

Applied with multicast transmission techniques in mobile computing environments, a mobile host will experience join and graft delay, happened when a host wants to join a multicast group in the fixed network, if there are no same multicast group member in the new cell the mobile host enters. Due to low bandwidth and higher error rate, there happens many additional traffic. In this paper, we propose a pre-join technique which new mobile support station joins the multicast group in advance based on signal strength hint in the current cell. We use the multiple level acknowledgement strategy that executes acknowledgment separately between the fixed part and the wireless transmission path. Using our strategy, it is an efficient technique in case there are more cells that has no multicast group members and less mobile host movements.

• KIPS Transactions on Computer and Communication Systems
• /
• v.9 no.10
• /
• pp.221-230
• /
• 2020

Supercomputer that shares limited resources to multiple users needs a way to optimize the execution of application. For this, it is useful for system administrators to get prior information and hint about the applications to be executed. In most high-performance computing system operations, system administrators strive to increase system productivity by receiving information about execution duration and resource requirements from users when executing tasks. They are also using profiling techniques that generates the necessary information using statistics such as system usage to increase system utilization. In a previous study, we have proposed a scheduling optimization technique by developing a hardware performance counter-based profiling technique that enables characterization of applications without further understanding of the source code. In this paper, we constructed a profiling testbed cluster to support optimal execution of the supercomputer and experimented with the scalability of the profiling method to analyze application characteristics in the built cluster environment. Also, we experimented that the profiling method can be utilized in actual scheduling optimization with scalability even if the application class is reduced or the number of nodes for profiling is minimized. Even though the number of nodes used for profiling was reduced to 1/4, the execution time of the application increased by 1.08% compared to profiling using all nodes, and the scheduling optimization performance improved by up to 37% compared to sequential execution. In addition, profiling by reducing the size of the problem resulted in a quarter of the cost of collecting profiling data and a performance improvement of up to 35%.