• Korean Journal of Cognitive Science
• /
• v.26 no.4
• /
• pp.377-392
• /
• 2015

Korean semantic role labeling (SRL) is usually performed by a machine learning and requires a lot of corpus. However, the Korean PropBank used in Korean SRL system is less than PropBank. It leads to a low performance. Therefore, we expand the annotated corpus and verb frames for Korean SRL system to expand the Korean PropBank corpus. Most of the SRL system have a domain-dependent performance so, the performance may decrease if domain was changed. In this paper, we use the domain adaptation technique to reduce decreasing performance with the existing corpus and the small size of new domain corpus. We apply the domain adaptation technique to Structural SVM and Deep Neural Network. The experimental result show the effectiveness of the domain adaptation technique.

• Journal of Broadcast Engineering
• /
• v.27 no.4
• /
• pp.511-518
• /
• 2022

Supervised learning based on deep learning has made a leap forward in various application fields. However, many supervised learning methods work under the common assumption that training and test data are extracted from the same distribution. If it deviates from this constraint, the deep learning network trained in the training domain is highly likely to deteriorate rapidly in the test domain due to the distribution difference between domains. Domain adaptation is a methodology of transfer learning that trains a deep learning network to make successful inferences in a label-poor test domain (i.e., target domain) based on learned knowledge of a labeled-rich training domain (i.e., source domain). In particular, the unsupervised domain adaptation technique deals with the domain adaptation problem by assuming that only image data without labels in the target domain can be accessed. In this paper, we explore the unsupervised domain adaptation techniques.

• ETRI Journal
• /
• v.36 no.3
• /
• pp.429-438
• /
• 2014

Semantic role labeling (SRL) is a task in natural-language processing with the aim of detecting predicates in the text, choosing their correct senses, identifying their associated arguments, and predicting the semantic roles of the arguments. Developing a high-performance SRL system for a domain requires manually annotated training data of large size in the same domain. However, such SRL training data of sufficient size is available only for a few domains. Constructing SRL training data for a new domain is very expensive. Therefore, domain adaptation in SRL can be regarded as an important problem. In this paper, we show that domain adaptation for SRL systems can achieve state-of-the-art performance when based on structural learning and exploiting a prior model approach. We provide experimental results with three different target domains showing that our method is effective even if training data of small size is available for the target domains. According to experimentations, our proposed method outperforms those of other research works by about 2% to 5% in F-score.

• Proceedings of the KSPS conference
• /
• 2005.04a
• /
• pp.53-56
• /
• 2005

In this paper, we propose a transformation based robust adaptation technique that uses the maximum mutual information(MMI) estimation for the objective function and the linear spectral transformation(LST) for adaptation. LST is an adaptation method that deals with environmental noises in the linear spectral domain, so that a small number of parameters can be used for fast adaptation. The proposed technique is called MMI-LST, and evaluated on TIMIT and FFMTIMIT corpora to show that it is advantageous when only a small amount of adaptation speech is used.

• Journal of KIISE
• /
• v.42 no.4
• /
• pp.475-482
• /
• 2015

Developing a high-performance Semantic Role Labeling (SRL) system for a domain requires manually annotated training data of large size in the same domain. However, such SRL training data of sufficient size is available only for a few domains. Performances of Korean SRL are degraded by almost 15% or more, when it is directly applied to another domain with relatively small training data. This paper proposes two techniques to minimize performance degradation in the domain transfer. First, a domain adaptation algorithm for Korean SRL is proposed which is based on the prior model that is one of domain adaptation paradigms. Secondly, we proposed to use simplified features related to morphological and syntactic tags, when using small-sized target domain data to suppress the problem of data sparseness. Other domain adaptation techniques were experimentally compared to our techniques in this paper, where news and Wikipedia were used as the sources and target domains, respectively. It was observed that the highest performance is achieved when our two techniques were applied together. In our system's performance, F1 score of 64.3% was considered to be 2.4~3.1% higher than the methods from other research.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.15 no.1
• /
• pp.99-107
• /
• 2019

Retinex-based image enhancement is a technique that utilizes the property that the human visual characteristics are sensitive to the difference from the surrounding pixel value rather than the pixel value itself. These Retinex-based algorithms show different characteristics of the improved image depending on the applied color space or gamma correction. In this paper, we set eight different experimental conditions according to the application of color space and gamma correction, and analyze the objective and subjective performance of each Retinex based image enhancement algorithm and apply it to the implementation of Retinex based algorithm. In the case of gamma correction, quantitative low entropy images and low contrast images are obtained. The application of Retinex technique in HSI color space rather than RGB color space is found to be high in overall subjective image quality as well as maintaining color.

• Journal of Communications and Networks
• /
• v.16 no.4
• /
• pp.447-457
• /
• 2014

In this paper, we investigate the frequency domain channel estimation for multiple-input multiple-output (MIMO) single-carrier frequency-division multiple-access (SC-FDMA) systems. In MIMO SC-FDMA, code-division multiplexed (CDM) pilots such as cyclic-shifted Zadoff-Chu sequences have been adopted for channel estimation. However, most frequency domain channel estimation schemes were developed based on frequency-division multiplexing of pilots. We first develop a channel estimation error model by using CDM pilots, and then analyze the mean-square error (MSE) of various minimum MSE (MMSE) frequency domain channel estimation techniques. We show that the cascaded one-dimensional robust MMSE (C1D-RMMSE) technique is complexity-efficient, but it suffers from performance degradation due to the channel correlation mismatch when compared to the two-dimensional MMSE (2D-MMSE) technique. To improve the performance of C1D-RMMSE, we design a robust iterative channel estimation (RITCE) with a frequency replacement (FR) algorithm. After deriving the MSE of iterative channel estimation, we optimize the FR algorithm in terms of the MSE. Then, a low-complexity adaptation method is proposed for practical MIMO SC-FDMA systems, wherein FR is performed according to the reliability of the data estimates. Simulation results show that the proposed RITCE technique effectively improves the performance of C1D-RMMSE, thus providing a better performance-complexity tradeoff than 2D-MMSE.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.30 no.2
• /
• pp.1-11
• /
• 2002

In the present study, a parallel DSCM technique based on a cell-based data structure is developed for the efficient simulation of rarefied gas flows especially od PC clusters. Dynamic load balancing is archieved by decomposing the computational domain into several sub-domains and accounting for the number of particles and the number cells of each domain. Mesh adaptation algorithm is also applied to improve the resolution of the solution and to reduce the grid dependency. It was demonstrated that accurate solutions can be obtained after several levels of mesh adapation starting from a coars initial grid. The method was applied to a two-dimensioanal supersonic leading-edge flow and the axi-symmetric Rothe nozzle flow to validate the efficiency of the present method. It was found that the present method is a very effective tool for the efficient simulation of rarefied gas flow on PC-based parallel machines.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1985.10a
• /
• pp.87-94
• /
• 1985

In this paper, enhancement of speech corrupted by additive white or colored noise is stuided. The nuconstrained frequency-domain block least-mean-square (UFBLMS) adaptation algorithm and its frequency-weighted version are newly applied to speech enhancement. For enhancement of speech degraded by white noise, the performance of the UFBLMS algorithm is superior to the spectral subtraction method or Wiener filtering technique by more than 3 dB in segmented frequency-weighted signal-to-noise ratio(FWSNERSEG) when SNR of speech is in the range of 0 to 10 dB. As for enhancement of noisy speech corrupted by colored noise, the UFBLMS algorithm is superior to that of the spectral subtraction method by about 3 to 5 dB in FWSNRSEG. Also, it yields better performance by about 2 dB in FWSNR and FWSNRSEG than that of time-domain least-mean-square (TLMS) adaptive prediction filter(APF). In view of the computational complexity and performance improvement in speech quality and intelligibility, the frequency-weighted UFBLMS algorithm appears to yield the best performance among various algorithms in enhancing noisy speech corrupted by white or colored noise.

• The Journal of Information Systems
• /
• v.8 no.1
• /
• pp.27-44
• /
• 1999

Case-Based Reasoning(CBR) offers a new approach for developing knowledge based systems. CBR has several research issues which can be divided into two categories : (1) static issues and (2) dynamic issues. The static issues are related to case representation scheme and case data model, that is, focus on casebase which is a repository of cases. The dynamic issues, on the other hand, are related to case retrieval procedure and problem solving process, i.e. case adaptation phase. This research is forcused on retrieval procedure Traditional query processing accepts precisely specified queries and only provides exact answers, thus requiring users to fully understand the problem domain and the casebase schema, but returning limited or even null information if the exact answer is not available. To remedy such a restriction, extending the classical notion of query answering to approximate query answering(AQA) has been explored. AQA can be achieved by neighborhood query answering or associative query answering. In this paper, neighborhood query answering technique is used for AQA. To reinforce the CBR process, a new retrieval procedure(cooperative CBR) using neighborhood query answering is proposed. An neighborhood query answering relaxes a query scope to enlarge the search range, or relaxes an answer scope to include additional information. Computer Aided Process Planning(CAPP) is selected as cooperative CBR application domain for test. CAPP is an essential key for achieving CIM. It is the bridge between CAD and CAM and translates the design information into manufacturing instructions. As a result of the test, it is approved that the problem solving ability of cooperative CBR is improved by relaxation technique.