http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
한국 정신장애의 역학 조사 연구[I] : 각 정신장애의 유병률
조맹제,함봉진,김장규,박강규,정은기,서동우,김선욱,조성진,이준영,홍진표,최용성,박종익,이동우,이기철,배재남,신정호,정인원,박종한,배안,이충경 大韓神經精神醫學會 2004 신경정신의학 Vol.43 No.4
Objectives : This study aims to estimate the prevalence of the DSM-IV psychiatric disorders in Korean population using the Korean version of Composite International Diagnostic Interview (K-CIDI). Methods : Subjects were selected by taking multi-stage, cluster samples of 7,867 adult household residents, 18 to 64 years of age, in ten catchment areas. Total 78 trained interviewers administered the K-CIDI to the selected respondents, from June 1 to November30,2001. Results : Total 6,275 respondents completed the interview. Some 33.5% of respondents reported at least one lifetime disorder, 20.6% reported at least one-year disorder, and 16.7% reported at least one-month disorder. The most common lifetime disorders were alcohol abuse/dependence (17.24%), nicotine dependence/withdrawal (11.19%), specific phobia (5.16%), and major de-pressive disorder (4.25%). The lifetime prevalence of substance abuse/dependence (0.25%) and schizophrenia (0.16%) was very low. Nicotine and alcohol use disorder showed very high male/female ratio. Mood disorder and anxiety disorder were more prevalent among female than male. Conclusion : The prevalence of psychiatric disorders was high. In comparison with other studies, remarkable differences in distributions of psychiatric disorders across the areas and times were observed.
On the Distributed Virtual Channel Based MAC for Spatial Reuse in Wireless LANs
( Choong Seon Hong ) 한국정보처리학회 2006 한국정보처리학회 학술대회논문집 Vol.13 No.2
Existing MAC protocols for wireless LANs employ CSMA/CA that avoids collisions by preserving the wireless medium exclusively within 2-hop network. This medium reservation obstructs neighboring nodes at both sides from concurrent transmission. This problem reduces medium utilization and overall throughput in Wireless LANs. In this work, we offer parallel data transfers using concurrent virtual channels over same physical channel by distributed transmit/receive synchronization. Each virtual channel is used by a transmitter/receiver pair and all virtual channels within 2-hop network utilize the medium and avoid inter virtual channel interference. This virtual channel based protocol eliminates the exclusive medium reservation and improves the performance in terms of latency and overall network throughput.
Probabilistic Exposure Identification for Wireless Sensor Network
( Choong Seon Hong ) 한국정보처리학회 2006 한국정보처리학회 학술대회논문집 Vol.13 No.2
Sensing and coverage are the two most relevant tasks for a sensor network. Wireless sensor network performance entirely depends on the success of sensing and detecting any object over a monitored region. Network performance may always degrade due to the limited energy and computation capability of sensors. So, it is a crucial task to identify those humiliations of sensors in a network. Exposure path is a probabilistic concept and it determines the probability of detecting any object by the sensors. In this paper we proposed a new algorithm to find out the minimum exposure path in a particular region of the network. It is totally a graph-theory as well as a convex geometry based approach. In order to improve the network performance the minimum exposure path algorithm of this paper can perform an appreciable role. Moreover the use of closest sensor of a cluster makes the overall computation less expensive for the proposed algorithm. Finally, we analyzed the proposed algorithm with some numerical simulation in order to obtain an approximate minimum exposure path.
Optimized Quantization for Convolutional Deep Neural Networks in Federated Learning
Choong Seon Hong,You Jun Kim 한국통신학회 2020 한국통신학회 APNOMS Vol.2020 No.09
Federated learning is a distributed learning method that trains a deep network on user devices without collecting data from central server. It is useful when the central server can’t collect data. However, the absence of data on central server means that deep network compression using data is not possible. Deep network compression is very important because it enables inference even on device with low capacity. In this paper, we proposed a new quantization method that significantly reduces FPROPS(floating-point operations per second) in deep networks without leaking user data in federated learning. Quantization parameters are trained by general learning loss, and updated simultaneously with weight. We call this method as OQFL(Optimized Quantization in Federated Learning). OQFL is a method of learning deep networks and quantization while maintaining security in a distributed network environment including edge computing. We introduce the OQFL method and simulate it in various Convolutional deep neural networks. We shows that OQFL is possible in most representative convolutional deep neural network. Surprisingly, OQFL(4bits) can preserve the accuracy of conventional federated learning(32bits) in test dataset.