• ETRI Journal
• /
• 제32권3호
• /
• pp.482-485
• /
• 2010

Feature selection is very important for feature-based relation classification tasks. While most of the existing works on feature selection rely on linguistic information acquired using parsers, this letter proposes new features, including probabilistic and semantic relatedness features, to manifest the relatedness between patterns and certain relation types in an explicit way. The impact of each feature set is evaluated using both a chi-square estimator and a performance evaluation. The experiments show that the impact of relatedness features is superior to existing well-known linguistic features, and the contribution of relatedness features cannot be substituted using other normally used linguistic feature sets.

• 한국언어정보학회지:언어와정보
• /
• 제14권2호
• /
• pp.47-70
• /
• 2010

This study investigates the semantic and prosodic properties of the so-called contrastive topic. We posit two informational primitives, namely, topical feature [+-T] and focal feature [+-F], from which four different informational categories, i.e., [+T, +F], [+T, -F], [-T, +F], and [-T, -F], are yielded. It is proposed that the informational category of contrastive topic has focal property [+F] as well as topical property [+T]. Based on the semantic approach that regards the function of [+F] as identificational predication and that of [+T] as forming a semantic conditional clause, it is shown that the semantic function of contrastive topic, which is specified as [+T, +F], is the combination of these two functions, i.e., identificational predication in a semantic conditional clause. This is supported by a scrutinized exploration of the prosodic pattern of English contrastive topic.

• Journal of Information Processing Systems
• /
• 제14권6호
• /
• pp.1318-1330
• /
• 2018

Video captioning refers to the process of extracting features from a video and generating video captions using the extracted features. This paper introduces a deep neural network model and its learning method for effective video captioning. In this study, visual features as well as semantic features, which effectively express the video, are also used. The visual features of the video are extracted using convolutional neural networks, such as C3D and ResNet, while the semantic features are extracted using a semantic feature extraction network proposed in this paper. Further, an attention-based caption generation network is proposed for effective generation of video captions using the extracted features. The performance and effectiveness of the proposed model is verified through various experiments using two large-scale video benchmarks such as the Microsoft Video Description (MSVD) and the Microsoft Research Video-To-Text (MSR-VTT).

• 한국멀티미디어학회논문지
• /
• 제24권1호
• /
• pp.1-12
• /
• 2021

In this paper, we present a new architecture for semantic segmentation. Semantic segmentation aims at a pixel-wise classification which is important to fully understand images. Previous semantic segmentation networks use features of multi-layers in the encoder to predict final results. However, they do not contain various receptive fields in the multi-layers features, which easily lead to inaccurate results for boundaries between different classes and small objects. To solve this problem, we propose a multi-path feature fusion module that allows for features of each layers to contain various receptive fields by use of a set of dilated convolutions with different dilatation rates. Various experiments demonstrate that our method outperforms previous methods in terms of mean intersection over unit (mIoU).

• 한국정보시스템학회지:정보시스템연구
• /
• 제31권4호
• /
• pp.01-23
• /
• 2022

Purpose This study aims to propose a deep learning-based fake review detection model by combining authors' linguistic features and semantic information of reviews. Design/methodology/approach This study used 358,071 review data of Yelp to develop fake review detection model. We employed linguistic inquiry and word count (LIWC) to extract 24 linguistic features of authors. Then we used deep learning architectures such as multilayer perceptron(MLP), long short-term memory(LSTM) and transformer to learn linguistic features and semantic features for fake review detection. Findings The results of our study show that detection models using both linguistic and semantic features outperformed other models using single type of features. In addition, this study confirmed that differences in linguistic features between fake reviewer and authentic reviewer are significant. That is, we found that linguistic features complement semantic information of reviews and further enhance predictive power of fake detection model.

• 산경연구논집
• /
• 제12권5호
• /
• pp.47-58
• /
• 2021

Purpose - Current research aimed at exploring the effect differences between the two types of processing product information such as the imagining and the considering on psychological product ownership which could influence the intent to purchase or use the product, and focused on identifying the interaction effects of activated memory information type and advertising information type on each of the information processing types. Research design, data, and methodology - This study divided the information processing types into imagining and considering, and the consumer's memories were divided into autobiographical or episodic and semantic memory. The advertising information was approached in each of event information being together with the product and product feature information. At empirical study, 2(two types of memory activation: episodic and semantic memory activation) ∗ 2(two types of advertising information: event-focused and product feature-focused advertising information) between-subjects design was used to make four types of questionnaire according to the type of experimental groups. Through the survey platform, 'questionnaire stars' of 'WeChat' in China, 219 questionnaire data were collected for empirical study. The structural equation model in AMOS 26 and Anova were used to verify hypotheses. Results - First, the ownership affected the usage intent positively. Second, the imagining did not affect the psychological ownership but did directly affect the usage intention, and the considering affected the ownership positively. Third, the episodic memory activation positively influenced the imagining and negatively affected the considering, whereas the semantic memory activation positively influenced the considering and negatively affected the imagining. Fourth, event-advertising information increased the effects of the activated episodic memory on the imagining, and feature-advertising information increased the effects of the activated semantic memory on the considering. Conclusions - marketers should develop and advertise their product-related event message to trigger the imaging that directly increase the intent to purchase or use their product, when consumers are under the activation of their episodic memory. And marketers should advertise their product feature-related message to trigger the considering that could induce consumers' ownership for their product to increase the intent to purchase or use their product, when they are under the activation of their semantic memory.

• 한국콘텐츠학회논문지
• /
• 제6권3호
• /
• pp.12-19
• /
• 2006

본 논문에서는 질의 응답 시스템의 성능 향상을 위한 질의문 심층 분석을 제안한다. 일반적인 질의응답 시스템들은 사용자의 자연언어 질의의 의미를 분석하지 않기 때문에 정확한 정답을 제공하는 것이 어렵다. 질의문 심층 분석은 의미자질 추출 문법과 자연언어 질의 특성을 이용하여 사용자의 질의를 의미적으로 분석하고, 의미자질들을 추출한다. 의미자질 추출 문법과 자연언어 질의 특성은 사용자 질의의 의미와 구문 구조를 반영하기 위해 의미자질과 형식형태소로 표현된다. 웹에서 추출한 세부 정답 유형이 '인물'인 100개의 질의에 대한 실험을 통해, 비교적 짧지만 사용자의 질의 의도를 충분히 표현하고 있는 자연언어 질의에 대해 질의문 심층 분석을 수행함으로써 사용자의 질의 의도를 분석하고, 의미자질들을 추출할 수 있음을 보였다.

• Journal of Multimedia Information System
• /
• 제5권2호
• /
• pp.67-78
• /
• 2018

This article proposes the modified KNN (K Nearest Neighbor) algorithm which considers the feature similarity and is applied to the word categorization. The texts which are given as features for encoding words into numerical vectors are semantic related entities, rather than independent ones, and the synergy effect between the word categorization and the text categorization is expected by combining both of them with each other. In this research, we define the similarity metric between two vectors, including the feature similarity, modify the KNN algorithm by replacing the exiting similarity metric by the proposed one, and apply it to the word categorization. The proposed KNN is empirically validated as the better approach in categorizing words in news articles and opinions. The significance of this research is to improve the classification performance by utilizing the feature similarities.

• 디지털융복합연구
• /
• 제19권6호
• /
• pp.251-258
• /
• 2021

본 논문에서는 U-Net 기반의 semantic segmentation 방법에서 정확도를 향상시키기 위해 residual learning을 활용한 인코더-디코더 구조의 모델을 제안하였다. U-Net은 딥러닝 기반의 semantic segmentation 방법이며 자율주행 자동차, 의료 영상 분석과 같은 응용 분야에서 주로 사용된다. 기존 U-Net은 인코더의 얕은 구조로 인해 특징 압축 과정에서 손실이 발생한다. 특징 손실은 객체의 클래스 분류에 필요한 context 정보 부족을 초래하고 segmentation 정확도를 감소시키는 문제가 있다. 이를 개선하기 위해 제안하는 방법은 기존 U-Net에 특징 손실과 기울기 소실 문제를 방지하는데 효과적인 residual learning을 활용한 인코더를 통해 context 정보를 효율적으로 추출하였다. 또한, 인코더에서 down-sampling 연산을 줄여 특징맵에 포함된 공간 정보의 손실을 개선하였다. 제안하는 방법은 Cityscapes 데이터셋 실험에서 기존 U-Net 방법에 비해 segmentation 결과가 약 12% 향상되었다.

• 융합정보논문지
• /
• 제11권10호
• /
• pp.45-52
• /
• 2021

본 논문에서는 U-Net 기반의 semantic segmentation 방법에서 정확도를 개선하기 위한 Atrous Residual U-Net (AR-UNet)을 제안하였다. U-Net은 의료 영상 분석, 자율주행 자동차, 원격 감지 영상 등의 분야에서 주로 사용된다. 기존 U-Net은 인코더 부분에서 컨볼루션 계층 수가 적어 추출되는 특징이 부족하다. 추출된 특징은 객체의 범주를 분류하는 데 필수적이며, 부족할 경우 분할 정확도를 저하시키는 문제를 초래한다. 따라서 이 문제를 개선하기 위해 인코더에 residual learning과 ASPP를 활용한 AR-UNet을 제안하였다. Residual learning은 특징 추출 능력을 개선하고, 연속적인 컨볼루션으로 발생하는 특징 손실과 기울기 소실 문제 방지에 효과적이다. 또한 ASPP는 특징맵의 해상도를 줄이지 않고 추가적인 특징 추출이 가능하다. 실험은 Cityscapes 데이터셋으로 AR-UNet의 효과를 검증하였다. 실험 결과는 AR-UNet이 기존 U-Net과 비교하여 향상된 분할 결과를 보였다. 이를 통해 AR-UNet은 정확도가 중요한 여러 응용 분야의 발전에 기여할 수 있다.