http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Deep learning can contrast the minimal pairs of syntactic data
Kwonsik Park,Myung-Kwan Park,Sanghoun Song 경희대학교 언어정보연구소 2021 언어연구 Vol.38 No.2
The present work aims to assess the feasibility of using deep learning as a useful tool to investigate syntactic phenomena. To this end, the present study concerns three research questions: (i) whether deep learning can detect syntactically inappropriate constructions, (ii) whether deep learning’s acceptability judgments are accountable, and (iii) whether deep learning’s aspects of acceptability judgments are similar to human judgments. As a proxy for a deep learning language model, this study chooses BERT. The current paper comprises syntactically contrasted pairs of English sentences which come from the three test suites already available. The first one is 196 grammatical -ungrammatical minimal pairs from DeKeyser (2000). The second one is examples in four published syntax textbooks excerpted from Warstadt et al. (2019). The last one is extracted from Sprouse et al. (2013), which collects the examples reported in a theoretical linguistics journal, Linguistic Inquiry. The BERT models, base BERT and large BERT, are assessed by judging acceptability of items in the test suites with an evaluation metric, surprisal, which is used to measure how ‘surprised’ a model is when encountering a word in a sequence of words, i.e., a sentence. The results are analyzed in the two frameworks: directionality and repulsion. The results of directionality reveals that the two versions of BERT are overall competent at distinguishing ungrammatical sentences from grammatical ones. The statistical results of both repulsion and directionality also reveal that the two variants of BERT do not differ significantly. Regarding repulsion, correct judgments and incorrect ones are significantly different. Additionally, the repulsion of the first test suite, which is excerpted from the items for testing learners’ grammaticality judgments, is higher than the other test suites, which are excerpted from the syntax textbooks and published literature. This study compares BERT’s acceptability judgments with magnitude estimation results reported in Sprouse et al. (2013) in order to examine if deep learning’s syntactic knowledge is akin to human knowledge. The error analyses on incorrectly judged items reveal that there are some syntactic constructions that the two BERTs have trouble learning, which indicates that BERT’s acceptability judgments are distributed not randomly.
Not Yet as Native as Native Speakers: Comparing Deep Learning Predictions and Human Judgments
Kwonsik Park,Seok-Hoon You,Sanghoun Song 한국영어학학회 2020 영어학연구 Vol.26 No.1
The purpose of this paper is to examine feasibility of replacing humans with deep learning in nativeness judgments and figure out in which way to develop the model in order to reach the level of humans by comparing nativeness judgments by deep learning and humans on English data. The controlled items, composed of 210 sentences, are categorized into two types: well-formedness test (i.e., no syntactic violation) and plausibility (i.e., no awkwardness) test items, most of which are excerpted from precedent linguistics literature. The deep learning model and five English native speakers are asked to classify the nativeness of the same stimulus sentences and the results reveal differences and similarities between them; although the overall performance of humans overwhelms that of deep learning, they are quite similar in judging plausibility items and learner data. The length of response time―hanging back from decision of nativeness―does not guarantee the accuracy, which means judging nativeness depends on something like intuition rather than deliberation.
최소대립 문장쌍을 활용한 한국어 사전학습모델의 통사 연구 활용 가능성 검증
박권식 ( Kwonsik Park ),김성태 ( Seongtae Kim ),송상헌 ( Sanghoun Song ) 한국언어정보학회 2021 언어와 정보 Vol.25 No.3
Syntactic studies make use of the minimally pairwise sentences as an argumentation tool, because the pairs allow us to pay attention to the constraints of interest. Likewise, it is helpful to use a set of minimal pairs in deep learning-based experiments for assessing the syntactic ability of neural language models. In this context, this study verifies whether the deep learning Korean model has the ability to properly distinguish the well-formed expressions and the corresponding ill-formed expressions. In the meanwhile, this study serves to examine the feasibility of the language resource constructed by the Korean government for deep learning architecture. The research is three-fold. First, we conducted an acceptability judgment testing to verify whether and how the language resource used in this study is indeed trustworthy. The results indicate that the judgments provided in the language resource converge with the judgments of our own experiment well enough. Second, we employed four Korean models such as mBERT, KoBERT, KR-BERT, KorBERT in order to evaluate how the language resource has a potentiality to predict the well-formedness of Korean expressions. The different models yield different results, the reason of which is fully discussed. Third, we made use of an independent test-set for evaluating the deep learning systems. It turns out that the results are still challenging, which implies that the current Korean models may have room for improvement to understand the syntactic phenomena.
Cost-Effective Model for Energy Saving in Super-Tall Building
Song, Kwonsik,Park, Moonseo,Lee, Hyun-Soo,Kim, Sooyoung,Shin, Jinho Korea Institute of Construction Engineering and Ma 2013 Journal of construction engineering and project ma Vol.3 No.3
In many urban cities, super-tall buildings have been being constructed around New York and Chicago as the center since 1930 to improve the efficiency of land use and respond to new residential type. In terms of energy consumption, super-tall buildings are classified as a top energy consumption building. Also, as time passed, the degradation of energy performance occurs in super-tall buildings like general things so that these cannot show the initial performance planned in the design phase. Accordingly, building owners need to make a plan to apply energy saving measures to existing building during the operation phase. In order to select energy saving measures, calculus-based methods and enumerative schemes have been typically used. However, these methods are time-consuming and previous studies which used these methods have problems with not considering the initial construction cost. Consequently, this study proposes a model for selecting an optimal combination of energy saving measures which derives maximum energy saving within allowable cost using genetic algorithms. As a contribution of this research, it would be expected that a model is utilized as one of the decision-making tools during the planning stage for energy saving.
Artificial Tactile Sensor Structure for Surface Topography Through Sliding
Shin, Kwonsik,Sim, Minkyung,Choi, Eunmin,Park, Hyunchul,Choi, Ji-Woong,Cho, Yuljae,Sohn, Jung Inn,Cha, Seung Nam,Jang, Jae Eun IEEE 2018 IEEE/ASME transactions on mechatronics Vol.23 No.6
<P>Tactile sensors mimicking the human sense of touch have been studied and various technologies for the sensing of external stimuli have been suggested as well. Humans detect certain external stimuli and become aware of related sensations, such as roughness or smoothness. Among the various physical parameters, surface information is the most informative type of perception to impart these sensations onto an electromechanical system, such as an android robot or a smart phone. Here, an array sensor, which uses a sliding method for the precise perception of surface information, such as shapes and structures, is demonstrated. The suggested array sensor design with the excellent dynamic response of a piezoelectric material results in enhanced spatial resolutions with sliding motions and detects variable sliding speeds. A soft material was employed to the sensor to enhance the capability of shape distinction. Color mapping was applied to translate surface patterns into visual images. The reconfigured surface information had high accuracy compared to actual information. The demonstrated sliding speed, pattern detection, and shape detection capabilities as well as the higher spatial resolutions allow the sensor to be utilized as an artificial tactile sensor.</P>
연구성과에 있어 수확체감의 법칙의 타당성 검증: 한국 정부출연연구소 사례
김권식 ( Kim Kwonsik ),박준영 ( Park Juneyoung ),이광훈 ( Lee Kwanghoon ) 단국대학교 사회과학연구소 2017 공공정책과 국정관리 Vol.11 No.2
수확체감의 법칙은 경제학계에서 중요하게 받아들여지고 있는 가설 중 하나이다. 하지만 동 법칙을 정부출연연구소를 대상으로 검증한 연구는 많지 않다. 이에 본 연구는 정부출연연구소에 투입되는 자본 및 노동 요소와 이로 인해 산출되는 연구성과의 관계에도 수확체감의 법칙이 적용될 수 있는지를 검증한다. 이를 위해 한국의 정부출연연구소 평가결과를 패널자료로 구축하여 분석한 결과, 연구성과에 있어 수확체감의 가설이 채택되었다. 이러한 연구결과는 정부출연연구소의 연구성과를 향상시킬 수 있는 자본 및 노동 투입요소의 최적배분이 존재할 수 있음을 시사한다. The law of diminishing returns is a key ingredient of various models in economic literature. Though it is one of the most widely known economic hypotheses, few studies have to date been tested its validity in the area of government-funded R&D. This paper thus aims to test the law of diminishing returns of capital and labor factors in the research performance, using a panel data set taken from the performance evaluation results of Korean governmentfunded research institutes. Though the performance of the research institutes is our dependent variable measured by the government, the empirical evidence is consistent with the hypothesis that the law is valid for government-funded research institutes. Results may provide practical implications for the optimal allocation of resources to improve the performance of the research institutes.
웹 GIS기반의 건축문화재 홍수위험관리 시스템 개발에 관한 연구
전정호(Jeon, Jung-Ho),이현수(Lee, Hyun-Soo),박문서(Park, Moonseo),김현수(Kim, Hyunsoo),송권식(Song, KwonSik) 대한건축학회 2015 大韓建築學會論文集 : 構造系 Vol.31 No.1
In recent years, flood damage is drastically increasing due to global warming, urbanization, irregular weather condition and so on. Multilateral efforts are put into solving the issue, however, weakness in effectively responding to the flood risk toward the cultural heritage buildings located all over the nation exists. To address these limitations, this research attempts to present an effective flood management system by integrating Web GIS with Relational Database Management System(RDBMS) and using real-time rainfall data. Compared to the traditional system, we believe that this Web GIS based system is more efficient, adaptable and flexible.