• Journal of Korean Society for Atmospheric Environment
• /
• v.12 no.1
• /
• pp.29-41
• /
• 1996

The two-stage numerical model was used to study the relation between three-dimensional local wind model, advection/diffusion model of random walk method and second moment method on Pusan coastal area. The first stage is three dimensional time-dependent local wind model which gives the wind field and vertical dirrusion coefficient. The second stage is advection/diffusion model which uses the results of the first stage as input data. First, wind fields on Pusan coastal area for none synoptic scale wind showed typical land and sea breeze circulation, and convergence zone occured at 1200LST in northern of domain, in succession, moved northward of domain. Emissions from Sinpyeong industrial district were trasnported toward the inland by sea breeze during daytime, and reached the end part of domain about 1800LST. During nighttime, emissions return to sea by land breeze and vertical diffusion also contributes to upward transport. In order to use this model for forecast of air pollution concentration on the Pusan coastal area, it is necessary that computed value must be compared with measured value and wind fields model must also be dealt in detail.

• Journal of Intelligence and Information Systems
• /
• v.28 no.4
• /
• pp.251-273
• /
• 2022

Recently, research on applying text analysis to deep learning has steadily continued. In particular, researches have been actively conducted to understand the meaning of words and perform tasks such as summarization and sentiment classification through a pre-trained language model that learns large datasets. However, existing pre-trained language models show limitations in that they do not understand specific domains well. Therefore, in recent years, the flow of research has shifted toward creating a language model specialized for a particular domain. Domain-specific pre-trained language models allow the model to understand the knowledge of a particular domain better and reveal performance improvements on various tasks in the field. However, domain-specific further pre-training is expensive to acquire corpus data of the target domain. Furthermore, many cases have reported that performance improvement after further pre-training is insignificant in some domains. As such, it is difficult to decide to develop a domain-specific pre-trained language model, while it is not clear whether the performance will be improved dramatically. In this paper, we present a way to proactively check the expected performance improvement by further pre-training in a domain before actually performing further pre-training. Specifically, after selecting three domains, we measured the increase in classification accuracy through further pre-training in each domain. We also developed and presented new indicators to estimate the specificity of the domain based on the normalized frequency of the keywords used in each domain. Finally, we conducted classification using a pre-trained language model and a domain-specific pre-trained language model of three domains. As a result, we confirmed that the higher the domain specificity index, the higher the performance improvement through further pre-training.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.4
• /
• pp.1630-1642
• /
• 2016

Inter-domain routing is the most critical function of the Internet. The routing system is a logical network relying on the physical infrastructure with geographical characteristics. Nature disasters or disruptive accidents such as earthquakes, cable cuts and power outages could cause regional failures which fail down geographically co-located network nodes and links, therefore, affect the resilience of inter-domain routing system. This paper presents a model for regional failures in inter-domain routing system called REFER for the first time. Based on REFER, the resilience of the inter-domain routing system could be evaluated on a finer level of the Internet, considering different routing policies of intra-domain and inter-domain routing systems. Under this model, we perform simulations on an empirical topology of the Internet with geographical characteristics to simulate a regional failure locating at a city with important IXP (Internet eXchange Point). Results indicate that the Internet is robust under a city-level regional failure. The reachability is almost the same after the failure, and the reroutings occur at the edge of the Internet, hardly affecting the core of inter-domain routing system.

• Bulletin of the Korean Mathematical Society
• /
• v.46 no.6
• /
• pp.1237-1248
• /
• 2009

The domain of attraction of a nonlinear differential equations is the region of initial points of solution tending to the equilibrium points of the systems as the time going. Determining the domain of attraction is one of the most important problems to investigate nonlinear dynamical systems. In this article, we first present two algorithms to determine the domain of attraction by using the moment matrices. In addition, as an application we consider a class of SIRS infection model and discuss asymptotical stability by Lyapunov method, and also estimate the domain of attraction by using the algorithms.

• The KIPS Transactions:PartD
• /
• v.9D no.4
• /
• pp.659-666
• /
• 2002

Park, Seo and Kim (12) developed the input domain-based SRGM, which was able to quantitatively assess the reliability of a software system during the testing and operational phases. They assumed perfect debugging during testing and debugging phase. To make this input domain-based SRGM more realistic, this assumption should be relaxed. In this paper we generalize the input domain-based SRGM under imperfect debugging. Then its statistical characteristics are investigated.

• ETRI Journal
• /
• v.38 no.3
• /
• pp.487-493
• /
• 2016

Two new methods are proposed for an unsupervised adaptation of a language model (LM) with a single sentence for automatic transcription tasks. At the training phase, training documents are clustered by a method known as Latent Dirichlet allocation (LDA), and then a domain-specific LM is trained for each cluster. At the test phase, an adapted LM is presented as a linear mixture of the now trained domain-specific LMs. Unlike previous adaptation methods, the proposed methods fully utilize a trained LDA model for the estimation of weight values, which are then to be assigned to the now trained domain-specific LMs; therefore, the clustering and weight-estimation algorithms of the trained LDA model are reliable. For the continuous speech recognition benchmark tests, the proposed methods outperform other unsupervised LM adaptation methods based on latent semantic analysis, non-negative matrix factorization, and LDA with n-gram counting.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.2E
• /
• pp.51-55
• /
• 2004

In this paper, we propose a style-specific language model adaptation scheme using n-gram based tf*idf similarity for Korean spontaneous speech recognition. Korean spontaneous speech shows especially different style-specific characteristics such as filled pauses, word omission, and contraction, which are related to function words and depend on preceding or following words. To reflect these style-specific characteristics and overcome insufficient data for training language model, we estimate in-domain dependent n-gram model by relevance weighting of out-of-domain text data according to their n-. gram based tf*idf similarity, in which in-domain language model include disfluency model. Recognition results show that n-gram based tf*idf similarity weighting effectively reflects style difference.

• Journal of Institute of Control, Robotics and Systems
• /
• v.14 no.2
• /
• pp.111-117
• /
• 2008

This paper presents a new continuation-based quasi-steady-state(CQSS) time-domain simulation algorithm incorporating a multiplicative aggregated load model for power systems. The authors' previous paper introduced a CQSS algorithm, which has the robust convergent characteristic near the singularity point due to the application of a continuation method. The previous CQSS algorithm implemented the load restoration in power systems using the exponent-based load recovery model that is derived from the additive dynamic load model. However, the reformulated exponent-based model causes the inappropriate variation of short-term load characteristics when switching actions occur, during time-domain simulation. This paper depicts how to incorporate a multiplicative load restoration model, which does not have the problem of deforming short-term load characteristics, into the time simulation algorithm, and shows an illustrative example with a 39-bus test system.

• Journal of Communications and Networks
• /
• v.3 no.4
• /
• pp.342-350
• /
• 2001

The Internet routing instability, or the rapid fluctuation of network reachability information, is an important problem currently facing the Internet engineering community. High levels of network instability can lead to packet loss, increased network latency, and delayed routing convergence. At the extreme, high levels of routing instability can lead to the loss of internal connectivity in wide-area networks. In this paper, we investigate the variation of domain degree and domain count of the inter-domain network over time by using linear regression model in order to analyze the topology variation of inter-domain network. We Also propose an efficient policy management model to reduce the instability in the inter-domain routing system. The proposed model can be used to identify whether a routing policy is adequate to reduce convergence time that is required to return to a normal state when BGP routing instability happens. Experimental analysis shows that the proposed model can be used to set up routing policy in domains for the purpose of minimizing the effects and the propagation of BGP routing instability.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.6
• /
• pp.1599-1618
• /
• 2024

Access control has always been one of the effective methods to protect data security. However, in new computing environments such as big data, data resources have the characteristics of distributed cross-domain sharing, massive and dynamic. Traditional access control mechanisms are difficult to meet the security needs. This paper proposes CACM-MMSR to solve distributed cross-domain access control problem for massive resources. The method uses blockchain and smart contracts as a link between different security domains. A permission decision model migration method based on access control logs is designed. It can realize the migration of historical policy to solve the problems of access control heterogeneity among different security domains and the updating of the old and new policies in the same security domain. Meanwhile, a semantic reasoning-based permission decision method for unstructured text data is designed. It can achieve a flexible permission decision by similarity thresholding. Experimental results show that the proposed method can reduce the decision time cost of distributed access control to less than 28.7% of a single node. The permission decision model migration method has a high decision accuracy of 97.4%. The semantic reasoning-based permission decision method is optimal to other reference methods in vectorization and index time cost.