Earticle

Translators encounter tough problems in song translation: [1] the translated text should be as singable as the original lyric, and [2] the translated song should arouse the same deep emotion as the original music. The former condition imposes a number of formal constraints on song translation: e.g. matching syllables with musical notes, observing the beat and the rhythm of the original music, etc. The latter condition dictates functional requirements for translating songs. The goal of this paper is to figure out crucial factors that determine the emotion of the original song, and to propose appropriate translation strategies that help preserve the emotion in the target language. For this, we hypothesize that syntactic and prosodic parallel structures should contribute to the emotion of the song. Then, we investigate how these structures are translated by professional translators into singable Korean lyrics when they are asked to translate musical numbers in the English language. After all, we have found out that the translated texts do not coincide among translators when ST includes both syntactic and prosodic parallel structures, and that prosodic parallel structures are more difficult to translate than syntactic parallel structures. Based on this, we discuss practical methods of preserving the emotion of the original song in the translated language.

인공지능은 다양한 분야에서 사람을 대신하여 성과를 보여준다. 게임 분야에서도 인공지능은 다양한 방면 으로 사용된다. NPC(Non Player Character) 제작, 게임 난이도, 콘텐츠 생성 등의 분야에서 활용되고 있다. 그 러나 게임에서 사용되는 인공지능은 우수한 성과를 보여주며 사람을 이기기 위해 연구된다. 우수한 인공지능 모델이 게임에 존재하여 게임의 난이도가 상승하게 되면 플레이어가 게임에 흥미를 잃는 경우도 있다. 따라 서 본 논문의 목적은 사람과의 대결이 아닌 사람을 서포트 하는 인공지능의 개발이다. 현재 모바일 시장의 게임에는 자동 사냥 시스템이 존재한다. 게임에서 이 시스템은 난수 방식을 사용하기 때문에 효율이 떨어진 다. 그러므로 본 논문에서는 병렬구조의 DNN 모델을 사용하여 효율적인 자동 사냥 시스템을 구축한다. 새로 구축한 모델과 기존에 사용하던 난수 방식의 모델을 게임을 통해 비교한다.

AI shows performance on behalf of people in various fields. In games, AI is used in many ways. It is used in fields such as NPC (Non Player Character) production, game difficulty, and content creation. However, the AI used in the game shows excellent performance and is researched to beat people. When a good AI model exists in a game, increasing the difficulty of the game, the player may lose interest in the game. Therefore, the purpose of this paper is to develop artificial intelligence that supports people, not confrontation with people. Currently, there is an auto play system for games in the mobile market. In games, this system is not efficient because it uses a random method. Therefore, in this paper, we build an efficient auto play system using parallel DNN model. The newly constructed model and the existing random number model are compared through the game.

This study aims at analysing the syntactic characteristics of perception verb constructions in Old French(OF) based on two texts “La Chanson de Roland(1090; CDR)” and “Le Chevalier de la Charrete(1179-1181; Ch)”, and observing the evolution and change of their V2-structure. We will sort out and analyse the structural distribution of these constructions from the modern syntactic perspective while comparing them with those of Modern French(MF). According to the results of comparative analysis between the two texts, we will verify which text complies with MF syntactic structure in a more precise manner. Finally, apart from the occurrences of null subjects in OF constructions, we will confirm that they are undergoing the development from V2 structure in order to form MF fundamental syntactic structures. We will also suggest the possibility that OF perception verb constructions can be accounted for by concepts of feature inheritance and parallel movement introduced in Chomsky(2008).

본 연구는 한국어와 인도네시아어의 정보구조(information structure)를 비교·대조하 여 공통점과 차이점을 밝히는 데 그 목적이 있다. 한국어와 인도네시아어 드라마 병렬말 뭉치를 구축한 후 이를 바탕으로 두 언어의 구어 정보구조의 특징을 분석한다. 본 연구 에서 사용하는 정보구조의 개념은 ‘주어짐(givenness)’으로 화자와 청자가 이미 알고 있 는지의 사실 여부에 따라 신정보(new information)와 구정보(given information)로 분류된다. 연구 결과에 따르면, 신정보는 한국어에서는 상황에 따라 주격조사 ‘이/가’, 부정 (indefiniteness) 표현으로 표시되는 반면 인도네시아어에서는 어순과 보문소 ‘yang’으 로 표시된다. 구정보는 한국어에서는 보조사 ‘은/는’, 대명사, 용언 ‘있잖아/말이야’로 표시되고, 인도네시아어에서는 대명사와 접어 ‘=nya’로 표시된다. 구정보의 경우 한국 어와 인도네시아어 두 언어 모두에서 맥락을 통해 쉽게 드러날 수 있는 경우 해당하는 문장 성분을 생략할 수 있다. 또 한 가지 공통점은 수사구의 경우 핵 명사가 구정보이 면 이를 생략할 수 있는 있다는 점이다. 본 연구는 한국어와 인도네시아어의 구어 정보구조 대조에 관한 최초의 연구라는 데 그 의의가 있다. 본 연구의 결과는 인도네시아인 학습자를 위한 한국어 교육과 한국인 학습자를 위한 인도네시아어 교육의 기초 자료로 활용될 수 있을 것으로 기대한다.

The aim of this study is to reveal the commonalities and differences in Korean and Indonesian by comparing and contrasting the two languages’ spoken language information structures based on a parallel corpus. The notion of information structure used in this study is givenness, which is classified into new and given information according to what the speaker already knows. The results of the study show that new information in Korean is indicated by the subjective case marker ‘이/가’ and indefinite expressions, while in Indonesian, it is indicated by word order and the complementizer ‘yang’. Given information in Korean is indicated by the topic marker ‘은/는’, pronouns, and the terms ‘있 잖아/말이야’, while in Indonesian, it is indicated by pronouns and the clitic ‘-nya’. In the case of given information, the constituent can be ellipted in both Korean and Indonesian. Both languages also show that in numeral phrases, if the noun is a given information, it can be ellipted. This study is meaningful in that it is the first study on Korean and Indonesian spoken language information structure. It is hoped that this study will serve as useful material for Indonesian Korean learners in the field of Korean language education.

본 연구는 한국어와 인도네시아어의 정보구조(information structure)를 비교·대조하 여 공통점과 차이점을 밝히는 데 그 목적이 있다. 한국어와 인도네시아어 드라마 병렬말 뭉치를 구축한 후 이를 바탕으로 두 언어의 구어 정보구조의 특징을 분석한다. 본 연구 에서 사용하는 정보구조의 개념은 ‘주어짐(givenness)’으로 화자와 청자가 이미 알고 있 는지의 사실 여부에 따라 신정보(new information)와 구정보(given information)로 분류된다. 연구 결과에 따르면, 신정보는 한국어에서는 상황에 따라 주격조사 ‘이/가’, 부정 (indefiniteness) 표현으로 표시되는 반면 인도네시아어에서는 어순과 보문소 ‘yang’으 로 표시된다. 구정보는 한국어에서는 보조사 ‘은/는’, 대명사, 용언 ‘있잖아/말이야’로 표시되고, 인도네시아어에서는 대명사와 접어 ‘=nya’로 표시된다. 구정보의 경우 한국 어와 인도네시아어 두 언어 모두에서 맥락을 통해 쉽게 드러날 수 있는 경우 해당하는 문장 성분을 생략할 수 있다. 또 한 가지 공통점은 수사구의 경우 핵 명사가 구정보이 면 이를 생략할 수 있는 있다는 점이다. 본 연구는 한국어와 인도네시아어의 구어 정보구조 대조에 관한 최초의 연구라는 데 그 의의가 있다. 본 연구의 결과는 인도네시아인 학습자를 위한 한국어 교육과 한국인 학습자를 위한 인도네시아어 교육의 기초 자료로 활용될 수 있을 것으로 기대한다.

The aim of this study is to reveal the commonalities and differences in Korean and Indonesian by comparing and contrasting the two languages’ spoken language information structures based on a parallel corpus. The notion of information structure used in this study is givenness, which is classified into new and given information according to what the speaker already knows. The results of the study show that new information in Korean is indicated by the subjective case marker ‘이/가’ and indefinite expressions, while in Indonesian, it is indicated by word order and the complementizer ‘yang’. Given information in Korean is indicated by the topic marker ‘은/는’, pronouns, and the terms ‘있 잖아/말이야’, while in Indonesian, it is indicated by pronouns and the clitic ‘-nya’. In the case of given information, the constituent can be ellipted in both Korean and Indonesian. Both languages also show that in numeral phrases, if the noun is a given information, it can be ellipted. This study is meaningful in that it is the first study on Korean and Indonesian spoken language information structure. It is hoped that this study will serve as useful material for Indonesian Korean learners in the field of Korean language education.

《詩經》의 동식물, 자연 현상, 인간 생활의 기물 등은 ‘경물’이라는 용어로 포괄할 수 있다. 경물은 주요 경물과 보조 경물로 나뉜다. 주요 경물은 한 작품의 전체 내용을 총괄하며 일정한 분위기를 조성하고, 보조 경물은 주요 경물이 형상화하는 주제를 뒷받침하면서 해당 장의 내용을 구성한다. 본고에서는 《詩經》에서 짝을 찾고 그리워하는 내용이 풀 캐는 모습으로 형상화된 양상을 주요­보조 구조 및 병렬 구조로 나누었다. 또한 풀 캐기 경물과 다른 경물의 관계 및 이들 경물이 각 구조에서 발휘하는 효과를 분석하였다.

Shijing[詩經] has a lot of motifs in each poem, and motifs have close associations one another in a poem. ‘Main motif’ shapes the main idea with combination of images to help the audience realize the content of the poem instantly. ‘Sub-motif’ assists the main motif to embody the content in detail, and composes the imagery of each stanza. The plant-picking motif is generally known to be related with themes of abandoned wife or seeking love. In this thesis we can find structures of the motifs in poems which have plant-picking motifs materializing the antithetical subjects of abandoned wife and seeking love. The property of each motif has variation depending of the main idea.

디지털 통신 시스템에서 채널에서 발생되는 ISI(Inter-Symbol Interference)를 줄이기 위해 사용되는 적응 등화기의 알고리즘으로 MCMA(Constant Modulus Algorithm)가 있다. MCMA는 비교적 간단한 연산량을 갖지만 어느 정도 적당한 수렴율과 정상 상태의 MSE(mean square error)를 가지므로 본 논문에서는 이를 개선키 위해 병렬 구조를 갖는 등화 방식을 제안 하며, MCMA(Modified Constant Modulus Algorithm)과 MDD(Modified Decision Directed Algorithm)으로 구성이 되어 있다. 제안 방식에서는 4-QAM 좌표와 16-QAM 좌표의 관계를 이용한, 새로운 비용 함수를 정의 하고, 만약 수신된 신호에 오프셋이 발생할 때, MCMA의 성능이 저하되므로 이를 극복하기 위해서 본 논문에서는 MDD(Modified Decision Direct) 알고리즘과 결합한 병렬형 구조를 적용하므로서 기존의 MCMA 보다 빠른 수렴 속도와 낮은 MSE를 갖는 개선된 성능을 얻을 수 있음을 컴퓨터 시뮬레이션으로 확인하였다.

In digital communication system that the Modified Constant Modulus Algorithm (MCMA) reduced the use of the adaptive equalization algorithm to combat the Inter-symbol Interference (ISI). MCMA is relatively brief operation. The major point of MCMA that it only achieves moderate convergence rate and steady state mean square error (MSE). In this paper suggest, MCMA equaliztion improve the performance with parallel structure. It combines Modified Constant Modulus Algorithm(MCMA) and Modified Decision Directed(MDD) algorithm. By exploiting the inherent structural relationship between the 4-QAM signal's coordinates and 16-QAM signal's coordinates, another style of cost function for Modified Constant Modulus Algorithm(MCMA) is defined and If it happen to offset of received signals and MCMA is poor performance in order to overcome this because the paper combines apply for MCMA and MDD(Modified Decision Direct) alrogthm. By computer simulation, we confirmed that the proposed PMCMA-MDD algorithm has the fater convergence rate and steady mean square error than the conventional MCMA.

본 논문에서는 시분산 채널에서 발생되는 진폭과 위상 찌그러짐을 보상하기위한 적응 등화 알고리즘에서 MCMA와 MDD를 병렬 구조로 구현한 PMCMA-MDD (Paralled Modified Constant Modulus Algorithm-Modified Decision Directed )와 단일 구조로 구현한 DW-MCMA (Dual Weight-Modified Constant Modulus Algorithm)의 성능을 비교, 평가하였다. 이들 알고리즘은 기본적으로 MCMA와 RCA의 축소 신호점을 이용하는 비용 함수를 적용하여 빠른 수렴 속도와 적은 잔류 isi를 얻을 수 있다. 구조적으로 PMCMA-MDD 알고리즘은 2개의 독립적인 등화 필터를 병렬로 사용하며 DW-MCMA 알고리즘은 1개의 등화 필터를 이용하여 2번에 걸쳐 탭 계수를 갱신한다. 2가지 알고리즘에 대한 컴퓨터 시뮬레이션을 통해 수신측에서 복원된 신호 성상도, 잔류 isi 특성 곡선에의한 수렴 시간과 수렴 속도, 정규화 MSE 및 주어진 시분산 채널의 신호대 잡음비에서 SER (Symbol Error Rate)등의 성능을 비교하였으며, 시뮬레이션 결과 모든 성능에서 PMCMA-MDD가 DW-MCMA보다 우월함을 알 수 있었다.

This paper deals with the performance comparison and evaluation of blind adaptive equalizer, the PMCMA-MDD and DW-MCMA, that is used for compensation of the amplitude and phase distortion which occurs in the time dispersive channel. Basically, these algorithms are modification of MCMA cost fuction in order to obtain the fast convergence speed and reduced residual isi by taking the parallel and serial double structured and the combination of the concept of RCA for the updating the tap coefficient. We implements the algorithm of it and compare the recovered constellation, residual isi, MSE charateristics curve and SER in the signal to noise ratio given the time dispersive channel. As a result of simulation, the PMCMA-MDD has a good in recovered constellation than DW-MCMA. But in the SER, the DW-MCMA has a good than PMCMA-MDD.

일반적인 보통 등급의 지반(ground) 상 구조물의 안정성 확보를 위한 설계입력정수(input-parameters) 도출 방법들은 보편적으로 잘 알려져 있다. 연구지역과 같은 습곡된(folded) 변성암(metamorphic rocks) 지반은 엽리가 (foliation) 촘촘히 발달해 있고, 엽리에 평행한 소규모 단층들(faults)이 분포하고 있어 설계입력정수 도출을 위한 특 별한 조사방법 및 시험, 등이 요구된다. 수 mm 간격의 엽리가 발달한 변성암 지반은 엽리면 직접전단강도시험(direct shear test), 엽리의 배향(strike/dip)과 맵핑(mapping), 엽리의 지하 연속성 파악을 위한 시추조사, 변성암반의 암반 분류(rock mass rating), 등이 필요하다. 특정 엽리면을 따라 발달한 소규모의 엽리평행단층(foliation-parallel faults) 이 다수 발달한 지반은 선구조선 분석, 단층 추적을 위한 지표지질맵핑, 단층면 직접전단강도시험, 등이 필수적이다. 습곡지반은 지질구조구(structural domain) 분석, 불연속면의 평사투영(stereonet)해석, 습곡축을 따른 전기비저항탐 사, 등의 추가조사로 설계입력정수 도출이 합리적이다.

Methods for deriving design input-parameters to ensure the stability of a structure on a common ground are generally well known. Folded metamorphic rocks, such as the study area, are highly foliated and have small faults parallel to the foliation, resulting in special research methods and tests to derive design input parameters, Etc. are required. The metamorphic rock ground with foliation development of several mm intervals has a direct shear test on the foliation surface, the strike/dip mapping of the foliation, the boring investigation to determine the continuity of the foliation, and the rock mass rating of the metamorphic rock. etc. are required. In the case of a large number of small foliation-parallel faults developed along a specific foliation plane, it is essential to analyze the lineament, surface geologic mapping for fault tracing, and direct shear test. Folded ground requires additional geological-structural-domain analysis, discontinuity analysis of stereonet, electrical resistivity exploration along the fold axis, and so on.

본 논문에서는 낮은 삽입 손실을 가지는 광대역 DSPSL (double-sided parallel-stripline) - CPS (coplanar stripline) 전이구조를 제안하였다. 제안하는 전이구조는 자연스러운 전계분포의 필드 변환과 DSPSL과 CPS 사이의 임피던스 정합의 관점에서 설계되었다. 전이구조의 명확한 설계방법을 제시하여 바닥면 그라운드의 변화와 전이구조의 길이를 결정할 수 있게 하였다. 설계의 제작된 전이구조의 특성은 전이구조 당 6.2 ～ 18.2 GHz의 대역에서 0.7 dB의낮은 삽입 손실을 가지고 있으며, 6.2 ～ 30 GHz 이상의 대역에서 1.25 dB의 삽입 손실을 가짐을 확인하였다.

A high-performance wideband transition from double-sided parallel-stripline (DSPSL) to coplanar stripline (CPS) is proposed. This transition is designed by consideration of gradual field transformation and optimal impedance matching between DSPSL and CPS. Clear design guidelines of proposed transition are provided to determine the ground shape and the transition length. The fabricated transition exhibits less than 0.7 dB insertion loss per transition for frequencies from 6.2 to 18.2 GHz, and less than 1.25 dB insertion loss to over 30 GHz.

Let M be a real hypersurface with almost contact metric structure $({\phi},{\xi},{\eta},g)$ in a non flat complex space form $M_n(c)$. In this paper, we prove that if the structure Jacobi operator $R_{\xi}$ is ${\xi}$-parallel and the Ricci tensor S commutes with the structure operator $\phi$, then a real hypersurface in $M_n(c)$ is a Hopf hypersurface. Further, we characterize such Hopf hypersurface in $M_n(c)$.

In this paper, first we introduce the full expression of the Riemannian curvature tensor of a real hypersurface M in the complex quadric Q^m from the equation of Gauss and some important formulas for the structure Jacobi operator R_ξ and its derivatives ∇R_ξ under the Levi-Civita connection ∇ of M. Next we give a complete classification of Hopf real hypersurfaces with Reeb parallel structure Jacobi operator, ∇_ξR_ξ = 0, in the complex quadric Q^m for m ≥ 3. In addition, we also consider a new notion of 𝒞-parallel structure Jacobi operator of M and give a nonexistence theorem for Hopf real hypersurfaces with 𝒞-parallel structure Jacobi operator in Q^m, for m ≥ 3.

In this paper, a new PID fuzzy controller(FC) based on parallel operation of PI and PD fuzzy control is presented. First, two fuzzy rule bases are constructed by separating the linguistic control rule for PID FC into two parts : one is e-.DELTA.e part, and the other is .DELTAL.$^{2}$e-.DELTA.e part. And then two FCs employing these rule bases indivisually are synthesized and run in parallel. The incremental control input is determined by taking weighted mean of the outputs of two FCs. The proposed PID FC improves the transient response of the system and gives better performance than the conventional PI FC.

In relation to the generalized Tanaka-Webster connection, we consider a new notion of parallel structure Jacobi operator for real hypersurfaces in complex two-plane Grassmannians and prove the non-existence of real hypersurfaces in $G_2({\mathbb{C}}^{m+2})$ with generalized Tanaka-Webster parallel structure Jacobi operator.

Let M be a real hypersurface of a complex space form with almost contact metric structure $({\phi},{\xi},{\eta},g)$. In this paper, we study real hypersurfaces in a complex space form whose structure Jacobi operator $R_{\xi}=R({\cdot},{\xi}){\xi}$ is ${\xi}$-parallel. In particular, we prove that the condition ${\nabla}_{\xi}R_{\xi}=0$ characterize the homogeneous real hypersurfaces of type A in a complex: projective space $P_n{\mathbb{C}}$ or a complex hyperbolic space $H_n{\mathbb{C}}$ when $g({\nabla}_{\xi}{\xi},{\nabla}_{\xi}{\xi})$ is constant and not equal to -c/24 on M, where c is a constant holomorphic sectional curvature of a complex space form.

Generally, multidimensional data such as image and spatial data require large amount of storage space. There is a limit to store and manage those large amounts of data in single workstation. If we manage the data on parallel computing environment which is being actively researched these days, we can get highly improved performance. In this paper, we propose a parallel multidimensional index structure that exploits the parallelism of the parallel computing environment. The proposed index structure is nP(processor)-nxmD(disk) architecture which is the hybrid type of nP-nD and 1P-nD. Its node structure in-creases fan-out and reduces the height of an index. Also, a range search algorithm that maximizes I/O parallelism is devised, and it is applied to k-nearest neighbor queries. Through various experiments, it is shown that the proposed method outperforms other parallel index structures.

The stringent performance requirement for the high efficiency of routing protocols on the Internet can be satisfied by exploiting the threaded border gateway protocol (TBGP) on multi-cores, but the state-of-the-art TBGP performance is restricted by a mass of contentions when racing to access the routing table. To this end, the highly-efficient parallel access approach appears to be a promising solution to achieve ultra-high route processing speed. This study proposes a general routing table structure consisting of two-level tries for fast parallel access, and it presents a heuristic-based divide-and-recombine algorithm to solve a mass of contentions, thereby accelerating the parallel route updates of multi-threading and boosting the TBGP performance. As a projected TBGP, this study also modifies the table operations such as insert and lookup, and validates their correctness according to the behaviors of the traditional routing table. Our evaluations on a dual quad-core Xeon server show that the parallel access contentions decrease sharply by 92.5% versus the traditional routing table, and the maximal update time of a thread is reduced by 56.8 % on average with little overhead. The convergence time of update messages are improved by 49.7%.

Modified compact combline bandpass filters are proposed based on the miniaturized quarter-wave transmission line which is composed of the parallel coupled line and lumped capacitors. The electrical length of the parallel coupled line in a resonator, which determines the size of combline bandpass filters, is just $5^{\circ}$ or $7^{\circ}$, resulting in a compact circuit area. The designed combline bandpass filter also has a wide upper stopband by suppressing the spurious passbands, not moving. Measured results of two fabricated filters centered at 400MHz show good agreement with the theoretical predications.

In this paper, the electromagnetic characteristics of a grounded slab and a parallel-plate structure are analyzed by the Spline-type Divided-Difference Interpolation (SDDI) technique. The technique efficiently evaluates the MoM impedance matrix elements of the multifold spectral or spatial domain integrals or summation in integro differential equations. The numerical results of the proposed method agree well with those of the corresponding literatures.