• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.52-54
• /
• 2023

최근 인공지능 기술의 발전에 따라서 다양한 서비스에서 음성 인식을 활용한 서비스를 제공하면서 음성 인식에 대한 중요성이 증가하고 있다. 이 논문에서는 국내에서 많이 사용되고 있는 대표적인 인공지능 서비스 API 를 제공하는 구글, ETRI, 네이버에 대해서 감정 음성 관점에서 그 차이를 평가하였다. AI Hub 에서 제공하는 감성 대화 말뭉치 데이터 셋의 일부인 음성 테스트 데이터를 사용하여 평가한 결과 ETRI API 가 문자 오류율 (1.29%)과 단어 오류율(10.1%)의 성능 지표에 대해서 가장 우수한 음성 인식 성능을 보임을 확인하였다.

• Journal of Information Technology and Architecture
• /
• v.9 no.3
• /
• pp.333-344
• /
• 2012

This paper proposes a system which recommends movies using information from social network services containing personal interest and taste. Method for establishing data is as follows. The system gathers movies' information from web sites and user's information from social network services such as Facebook and twitter. The data from social network services is categorized into six steps of emotion level for more accurate processing following users' emotional states. Gathered data will be established into vector space model which is ideal for analyzing and deducing the information with the system which is suggested in this paper. The existing similarity measurement method for movie recommendation is presentation of vector information about emotion level and similarity measuring method on the coordinates using Cosine measure. The deducing method suggested in this paper is two-phase arithmetic operation as follows. First, using general cosine measurement, the system establishes movies list. Second, using similarity measurement, system decides recommendable movie list by vector operation from the coordinates. After Comparative Experimental Study on the previous recommendation systems and new one, it turned out the new system from this study is more helpful than existing systems.

• Journal of the Korea Computer Graphics Society
• /
• v.29 no.3
• /
• pp.33-43
• /
• 2023

Not only emotions affect physical sensations, but they also have an impact on physical movements. The responses to emotions vary depending on the type of emotional stimuli. However, research on the effects of emotional stimuli on the activation of bodily movements has not been rigorously examined, and these effects have not been investigated in Autonomous Emotion Recognition (AER) systems. In this study, we aimed to compare the emotional responses of 20 participants to three types of emotional stimuli (words, pictures, and videos) and investigate their activation or deactivation for the AER system. Our dependent measures included emotional responses, computer-based self-reporting methods, and bodily movements recorded using motion capture devices. The results suggested that video stimuli elicited higher levels of emotional movement, and emotional movement patterns were similar across different types of emotional stimuli for happiness, sadness, anger, and neutrality. Additionally, the findings indicated that bodily changes observed during video stimuli had the highest classification accuracy. These findings have implications for future research on the bodily changes elicited by emotional stimuli.

• The Journal of the Korea Contents Association
• /
• v.16 no.11
• /
• pp.72-82
• /
• 2016

Various studies have been done in medical service area but they have just focused on the examination of the relationships between cause and effect variables. This study, thus, empirically analyzed qualitative data regarding medical service problems using word cloud technique. The major results of the paper are as follows. The data reveal ten sources in medical service - forced treatment, excess inspection, misdiagnosis, carelessness, inexperienced service, waiting for emergency, reservation problem, unkindness, process problem, and inconvenience. Major words in the category of irrecoverable service failure are misdiagnosis, careless treatment, and inexperienced service whereas those in recoverable service failure are unkind attitude and negative experience in reservation system. Those who experienced a medical service problem are usually engaged in a public act and they make public protests and legal action against very severe problems. The conclusion of this study also suggests a summary, implication, and agenda of the research.

• The Journal of Society for e-Business Studies
• /
• v.21 no.3
• /
• pp.15-28
• /
• 2016

This paper proposes a system recommending spatial information what user wants with collecting and analyzing tweets around the user's location by using the GPS information acquired in mobile. This system has built an emotion dictionary and then derive the recommendation score of morphological analyzed tweets to provide not just simple information but recommendation through the emotion analysis information. The system also calculates distance between the recommended tweets and user's latitude-longitude coordinates and the results showed the close order. This paper evaluates the result of the emotion analysis in a total of 10 areas with two keyword 'Restaurants' and 'Performance.' In the result, the number of tweets containing the words positive or negative are 122 of the total 210. In addition, 65 tweets classified as positive or negative by analyzing emotions after a morphological analysis and only 46 tweets contained the meaning of the positive or negative actually. This result shows the system detected tweets containing the emotional element with recall of 38% and performed emotion analysis with precision of 71%.

• Korean Journal of Food Science and Technology
• /
• v.52 no.2
• /
• pp.172-176
• /
• 2020

Recently, it has been reported that the positive emotional responses from previously consumed food could be transferred to the new food, affecting the acceptance of the food. This study was conducted to develop emotional lexicons for evaluating consumers' emotional responses to the food. Focus group interviews were conducted using 15 Koreans and 23 Chinese consumers to elicit emotional terms for Korean food HMR products. Using 23 Chinese participants who did not participate in the previous interview, emotional terms were screened through discussions in an interview setting. An online survey among 50 Koreans and 50 Chinese was carried out to evaluate and verify the valence and arousal potential of the selected terms. Elicited emotional terms in these two countries had similar valence and arousal potentials. However, cross-cultural differences were also found, mostly in arousal potential. Therefore, interpretation should be done carefully when comparing emotional responses between Korean and Chinese subjects.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.5
• /
• pp.609-617
• /
• 2011

This paper proposes a personalized food recommendation system using user profiles auto-generated from Twitter. The user profiles are generated by extracting nouns from Twitter, and calculating emotional scores according to whether each noun is collocated with emotion words. Representative noun information for each food is constructed by analyzing web pages relevant to foods. Appropriate foods for users can be recommended by calculating similarities among the extracted resources. The proposed system has an advantage in that it can always recommend foods even if a user is a newcomer.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.380-382
• /
• 2019

GAN(Generative Adversarial Network)은 정해진 학습 데이터에서 정해진 생성자와 구분자가 서로 각각에게 적대적인 관계를 유지하며 동시에 서로에게 생산적인 관계를 유지하며 가능한 긍정적인 영향을 주며 학습하는 기계학습 분야이다. 전통적인 문장 생성은 단어의 통계적 분포를 기반으로 한 마르코프 결정 과정(Markov Decision Process)과 순환적 신경 모델(Recurrent Neural Network)을 사용하여 학습시킨다. 이러한 방법은 문장 생성과 같은 연속된 데이터를 기반으로 한 모델들의 표준 모델이 되었다. GAN은 표준모델이 존재하는 해당 분야에 새로운 모델로써 다양한 시도가 시도되고 있다. 하지만 이러한 모델의 시도에도 불구하고, 지금까지 해결하지 못하고 있는 다양한 문제점이 존재한다. 이 논문에서는 다음과 같은 두 가지 문제점에 집중하고자 한다. 첫째, Sequential 한 데이터 처리에 어려움을 겪는다. 둘째, 무작위로 생성하기 때문에 사용자가 원하는 데이터만 출력되지 않는다. 본 논문에서는 이러한 문제점을 해결하고자, 부분적인 정답 제공을 통한 조건별 생산적 적대 생성망을 설계하여 이 방법을 사용하여 해결하였다. 첫째, Sequence to Sequence 모델을 도입하여 Sequential한 데이터를 처리할 수 있도록 하여 원시적인 텍스트를 생성할 수 있게 하였다. 둘째, 부분적인 정답 제공을 통하여 문장의 생성 조건을 구분하였다. 결과적으로, 제안하는 기법들로 원시적인 감정 텍스트를 생성할 수 있었다.

• KIISE Transactions on Computing Practices
• /
• v.23 no.4
• /
• pp.268-273
• /
• 2017

Recently, deep learning encoder-based approach has been actively applied in the field of sentiment classification. However, Long Short-Term Memory network deep learning encoder, the commonly used architecture, lacks the quality of vector representation when the length of the documents is prolonged. In this study, for effective classification of the sentiment documents, we suggest the use of attention method-based deep learning encoder that generates document vector representation by weighted sum of the outputs of Long Short-Term Memory network based on importance. In addition, we propose methods to modify the attention method-based deep learning encoder to suit the sentiment classification field, which consist of a part that is to applied to window attention method and an attention weight adjustment part. In the window attention method part, the weights are obtained in the window units to effectively recognize feeling features that consist of more than one word. In the attention weight adjustment part, the learned weights are smoothened. Experimental results revealed that the performance of the proposed method outperformed Long Short-Term Memory network encoder, showing 89.67% in accuracy criteria.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.611-613
• /
• 2018

KoNLTK는 한국어와 관련된 다양한 언어자원과 언어처리 도구들을 파이썬 플랫폼에서 하나의 인터페이스 환경에서 제공하기 위한 언어처리 플랫폼이다. 형태소 분석기, 개체명 인식기, 의존 구조 파서 등 기초 분석 도구들과 단어 벡터, 감정 분석 등 응용 도구들을 제공하여 한국어 텍스트 분석이 필요한 연구자들의 편의성을 증대시킬 수 있다.