• Journal of Broadcast Engineering
• /
• v.26 no.7
• /
• pp.855-861
• /
• 2021

Neural Architecture Search (NAS) is cutting-edge technology in the machine learning community. NAS Without Training (NASWOT) recently has been proposed to tackle the high demand of computational resources in NAS by leveraging some indicators to predict the performance of architectures before training. The advantage of these indicators is that they do not require any training. Thus, NASWOT reduces the searching time and computational cost significantly. However, NASWOT only considers high-performing networks which does not guarantee a fast inference speed on hardware devices. In this paper, we propose a multi objectives reward function, which considers the network's latency and the predicted performance, and incorporate it into the Reinforcement Learning approach to search for the best networks with low latency. Unlike other methods, which use FLOPs to measure the latency that does not reflect the actual latency, we obtain the network's latency from the hardware NAS bench. We conduct extensive experiments on NAS-Bench-201 using CIFAR-10, CIFAR-100, and ImageNet-16-120 datasets, and show that the proposed method is capable of generating the best network under latency constrained without training subnetworks.

• The KIPS Transactions:PartC
• /
• v.16C no.3
• /
• pp.373-382
• /
• 2009

we present more efficient method of a medium access for real-time ubiquitous sensor networks. Proposed MAC protocol is like the randomized CSMA/CA protocol, but unlike previous legacy protocols, it does not use a time-varying contention window from which a node randomly picks a transmission slot. To reduce the latency for the delivery of event reports from sensor nodes, we carefully decide to select a fixed-size contention window with non-uniform probability distribution of transmitting in each slot. We show that the proposed method can offer up to several times latency reduction compared to legacy of IEEE 802.11 as the size of the sensor network scales up to 256 nodes using a widely-used network simulation package, NS-2. We finally show that proposed MAC scheme comes close to meet bounds on the best latency being achieved by a decentralized CSMA-based MAC protocol for real-time ubiquitous sensor networks which is sensitive to latency.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.1
• /
• pp.35-57
• /
• 2021

Fog computing aims to provide the solution of bandwidth, network latency and energy consumption problems of cloud computing. Likewise, management of data generated by healthcare IoT devices is one of the significant applications of fog computing. Huge amount of data is being generated by healthcare IoT devices and such types of data is required to be managed efficiently, with low latency, without failure, and with minimum energy consumption and low cost. Failures of task or node can cause more latency, maximum energy consumption and high cost. Thus, a failure free, cost efficient, and energy aware management and scheduling scheme for data generated by healthcare IoT devices not only improves the performance of the system but also saves the precious lives of patients because of due to minimum latency and provision of fault tolerance. Therefore, to address all such challenges with regard to data management and fault tolerance, we have presented a Fault Tolerant Data management (FTDM) scheme for healthcare IoT in fog computing. In FTDM, the data generated by healthcare IoT devices is efficiently organized and managed through well-defined components and steps. A two way fault-tolerant mechanism i.e., task-based fault-tolerance and node-based fault-tolerance, is provided in FTDM through which failure of tasks and nodes are managed. The paper considers energy consumption, execution cost, network usage, latency, and execution time as performance evaluation parameters. The simulation results show significantly improvements which are performed using iFogSim. Further, the simulation results show that the proposed FTDM strategy reduces energy consumption 3.97%, execution cost 5.09%, network usage 25.88%, latency 44.15% and execution time 48.89% as compared with existing Greedy Knapsack Scheduling (GKS) strategy. Moreover, it is worthwhile to mention that sometimes the patients are required to be treated remotely due to non-availability of facilities or due to some infectious diseases such as COVID-19. Thus, in such circumstances, the proposed strategy is significantly efficient.

• Journal of Physiology & Pathology in Korean Medicine
• /
• v.24 no.4
• /
• pp.598-601
• /
• 2010

This experiment was performed to investigate whether the combined-preparation of water extracts of germinated brown rice (WGR), water extracts of cultured mountain ginseng (WCG) and 70% ethanol extracts of Longanae Arillus (ELA) has hypnotic effects and/or enhances pentobarbital-induced sleep behaviors through the GABAergic system. The combined-preparation of WGR and WCG reduced sleep latency and prolonged sleep time induced by pentobarbital. ELA also reduced sleep latency and prolonged sleep time induced by pentobarbital. However, WGR or WCG itself did not induce sleep. The combined-preparation of WGR, WCG and ELA strongly reduced sleep latency and prolonged sleep time via chloride influx into primary cultured cerebellar granule cells. In conclusion, the combined-preparation of WGR, WCG and ELA augments pentobarbital-induced sleep behaviors through the modification of GABAergic system.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.26 no.6B
• /
• pp.727-735
• /
• 2001

본 논문은 8-VSB Advanced Digital TV용으로 사용할 수 있도록 ATSC(Advanced Television Standard Committee)의 규약을 만족시키도록 구현한 Reed Solomon 디코더에 대하여 기술한다. 구현된 RS Decoder는 공유된 Tree 구조의 Arithmetic 블록을 사용하여 종래의 기술보다 더 효율적인 연산기 구조를 제안하였으며 빠른 에러 탐지와 정정 시간으로 인한 FIFO의 사용갯수와 Latency Time을 크게 감소시킨 개선된 구조를 제안한다. 일반적으로 2N+A만큼의 Latency Time과 FIFO 개수를 N+A 만큼으로 감소시켰다. 이 RS 디코더는 Verilog HDL로 설계되었고 Synopsys Design Compiler에 의해 합성되었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10e
• /
• pp.301-303
• /
• 2002

One of the remarkable problems in PIM-DM(Protocol Independent Multicast-Dense Mode) is the latency time, increasing for some periods. The reason of this problem is proved to the confusion of flooding prune message and leave prune message. We propose a convincing solution to this confusion and we prove this by network simulation.

• The Journal of the Korea Contents Association
• /
• v.8 no.7
• /
• pp.58-65
• /
• 2008

The overlay multicast that has been recently presented as an alternative for the IP multicast has been getting much attention by the advancement of network techniques to enforce routing in application-levels. In this paper, we verified the efficiency of Internet broadcasting through comparing latency time of TBCP with mOBCP that have distributed Tree-first based spanning-tree structures in overlay multicast techniques. We also propose effective tree structures that minimize latency time and provide competent service by choosing optimum parent nodes early in the base of these two methods over the same time slot in multi-user connection. The results of our comparison showed that average latency time in mOBCP techniques declined by 36% from those of TBCP technique.

• Journal of Korea Multimedia Society
• /
• v.9 no.3
• /
• pp.323-332
• /
• 2006

Fair packet scheduling algorithms supporting quality-of-services of real-time multimedia applications can be classified into the following two schemes in terms of the reference time used in calculating the timestamp of arriving packet; the Finish-Time (FT) and Start-Time (ST) schemes. The FT scheme, used in most schedulers, that has the property of an inversely rate-proportional latency is suitable to support various delay bounds because it can adjust the latency of a flow with raising the flow's reserved rate. However, the scheme may incur some bandwidth loss due to excess rate reservation. Meanwhile, although the ST scheme does not suffer from the bandwidth loss, it is hard to support multiple delay bounds because of its latency property relying on the number of flows. This paper is devoted to propose a ST scheme based scheduler to effectively support multiple delay bounds and analyze its performance comparing to the FT scheme based scheduler. The comparison results show that the proposed scheduler gives better utilization by up to 50%.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.10a
• /
• pp.75-78
• /
• 2007

Mobile IP provides hand-held devices with mobility which allows the user to do work over the network. However, handover time due transfer between access routers causes network delays and data loss. L2Trigger Handover expects this handover to take place, and executes L3 handover before L2 handover takes place, thereby reducing overall handover latency, although it still is an issue since handover latency between AR is not completely eliminated in L2 trigger handover. In this paper took into consideration where MIH is used in MIPv4 and using MIH Table when handover is about to occur in MN(Mobile Node), thereby pre-fetching data needed by Handover. In this way, when the handover is estimated, it improves the init time that L2trigger had. Furthermore we can find that we can execute the handover with shorten init time in smaller and narrow overlap length

• Journal of KIISE:Computer Systems and Theory
• /
• v.31 no.10
• /
• pp.543-552
• /
• 2004

This paper proposes an energy optimization technique for latency and quality constrained video applications. It consists of two key techniques: frame-skipping technique and buffering technique. While buffering increases the slack time utilization at the OS level. frame skipping Increases the slack time itself at the application level, and both enhance the effectiveness of the dynamic voltage scaling technique. We use an H.263 encoder application as a test vehicle to which the proposed technique is applied. Experiments demonstrate that the proposed technique achieves noticeable energy reduction satisfying the given latency and video quality constraints.