• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권5호
• /
• pp.1396-1412
• /
• 2023

Conversation modeling is an important and challenging task in the field of natural language processing because it is a key component promoting the development of automated humanmachine conversation. Most recent research concerning conversation modeling focuses only on the current utterance (considered as the current question) to generate a response, and thus fails to capture the conversation's logic from its beginning. Some studies concatenate the current question with previous conversation sentences and use it as input for response generation. Another approach is to use an encoder to store all previous utterances. Each time a new question is encountered, the encoder is updated and used to generate the response. Our approach in this paper differs from previous studies in that we explicitly separate the encoding of the question from the encoding of its context. This results in different encoding models for the question and the context, capturing the specificity of each. In this way, we have access to the entire context when generating the response. To this end, we propose a deep neural network-based model, called the Context Model, to encode previous utterances' information and combine it with the current question. This approach satisfies the need for context information while keeping the different roles of the current question and its context separate while generating a response. We investigate two approaches for representing the context: Long short-term memory and Convolutional neural network. Experiments show that our Context Model outperforms a baseline model on both ConvAI2 Dataset and a collected dataset of conversational English.

• 한국IT서비스학회지
• /
• 제16권4호
• /
• pp.149-160
• /
• 2017

Sentiment analysis is the process of determining whether a piece of document, text or conversation is positive, negative, neural or other emotion. Sentiment analysis has been applied for several real-world applications, such as chatbot. In the last five years, the practical use of the chatbot has been prevailing in many field of industry. In the chatbot applications, to recognize the user emotion, sentiment analysis must be performed in advance in order to understand the intent of speakers. The specific emotion is more than describing positive or negative sentences. In light of this context, we propose deep learning models for conducting multi-class sentiment analysis for identifying speaker's emotion which is categorized to be joy, fear, guilt, sad, shame, disgust, and anger. Thus, we develop convolutional neural network (CNN), long short term memory (LSTM), and multi-layer neural network models, as deep neural networks models, for detecting emotion in a sentence. In addition, word embedding process was also applied in our research. In our experiments, we have found that long short term memory (LSTM) model performs best compared to convolutional neural networks and multi-layer neural networks. Moreover, we also show the practical applicability of the deep learning models to the sentiment analysis for chatbot.

• 디지털산업정보학회논문지
• /
• 제19권4호
• /
• pp.87-96
• /
• 2023

ChatGPT, a leader in generative AI, can use natural expressions like humans based on large-scale language models (LLM). The ability to grasp the context of the language and provide more specific answers by algorithms is excellent. It also has high-quality conversation capabilities that have significantly developed from past Chatbot services to the level of human conversation. In addition, it is expected to change the operation method of the tourism industry and improve the service by utilizing ChatGPT, a generative AI in the tourism sector. This study was conducted to explore ChatGPT trends and paradigms in tourism. The results of the study are as follows. First, keywords such as tourism, utilization, creation, technology, service, travel, holding, education, development, news, digital, future, and chatbot were widespread. Second, unlike other keywords, service, education, and Mokpo City data confirmed the results of a high degree of centrality. Third, due to CONCOR analysis, eight keyword clusters highly relevant to ChatGPT in the tourism sector emerged.

• 한국융합학회논문지
• /
• 제13권1호
• /
• pp.13-21
• /
• 2022

최근 딥러닝 기반의 자연어처리 연구는 사전 훈련된 언어 모델을 통해 대부분의 자연어처리 분야에서 우수한 성능을 보인다. 특히 오토인코더 (auto-encoder) 기반의 언어 모델은 다양한 한국어 이해 분야에서 뛰어난 성능과 쓰임을 증명하고 있다. 그러나 여전히 디코더 (decoder) 기반의 한국어 생성 모델은 간단한 문장 생성 과제에도 어려움을 겪고 있으며, 생성 모델이 가장 일반적으로 쓰이는 대화 분야에서의 세부 연구와 학습 가능한 데이터가 부족한 상황이다. 따라서 본 논문은 한국어 생성 모델을 위한 멀티턴 대화 데이터를 구축하고 전이 학습을 통해 생성 모델의 대화 능력을 개선하여 성능을 비교 분석한다. 또한, 검색 모델을 통해 외부 지식 정보에서 추천 응답 후보군을 추출하여 모델의 부족한 대화 생성 능력을 보완하는 방법을 제안한다.

• 한국지역사회생활과학회지
• /
• 제22권2호
• /
• pp.247-255
• /
• 2011

This study analyzed the demand for a community-based tourism site using a poisson model, a negative binominal model, a truncated poisson model and a truncated negative binominal model as count data models. For these reasons, questionnaire surveys were conducted into 5 community-based tourism sites in Chuncheon city with 406 tourists, and was analyzed using the STATA program. The fitness levels of four models were significant(p=0.0000) using a likelihood ratio test. The study results suggest that the demand of community-based tourism sites for visiting tourists was influenced by a pre-visiting experience, recognition of sustainable tourism, visitation of downtown, purchase of souvenir or farm produce, conversation with regional residents, regional harmony, preservation of natural resources and sex within the poisson and truncated poisson models. However, the variables of visitation of downtown, preservation of natural resources and sex were not significant within the negative binominal model and the visitation of downtown and preservation of natural resources were not significant within the truncated negative binominal model. The results of the visiting demand of community-based tourism sites can provide information for sustainable regional development strategies.

• Asia pacific journal of information systems
• /
• 제14권4호
• /
• pp.49-70
• /
• 2004

Multi-agent technology is being regarded as one of the promising technologies for today's supply chain management because of its desirable features such as autonomy, intelligence, and collaboration. This paper suggests a multi-agent system architecture with which companies can improve the efficiency of their supply chains by collaborative operation. Reflecting the practical difficulties of collaboration in complex supply chains, the architecture allows agent systems to share information with only neighboring companies for the coordinated operation. The suggested architecture is elaborated with a collaboration model based on Petri-net, conversation models for communication, and internal behavior models of each agent. A simulation experiment was performed for the evaluation of the suggested architecture. The result implies that when the estimation of market demand is higher than a certain level, the suggested architecture can be beneficial.

• Journal of Public Diplomacy
• /
• 제2권1호
• /
• pp.43-68
• /
• 2022

Multiple modes of communication on social media can contribute to public diplomacy in informing, conversing, and networking with members of foreign publics. However, manipulative behaviours on social media, prevalent especially in high tension contexts, create disruptions to authentic communication in what could be grey/black propaganda or information warfare. This study reviews existing literature about models of public diplomacy to guide an empirical study of China's communication in the #SouthChinaSea conversation on Twitter. It uses computational methods to identify, record, and analyze one-way, two-way, and network communication of China's actors. It employs manual qualitative research to determine the nature of China's actors. On that basis, it assesses China's Twitter communication in the issue against various models of public diplomacy.

• 농촌계획
• /
• 제17권1호
• /
• pp.59-65
• /
• 2011

Re-visitation have an effect on dependent variables of regional tourism demand model. This study focused on the re-visitation impacts of consideration on sustainable tourism development of tourists as a new factors of tourism. Based on literature reviews, 11 variables were selected, a questionnaire survey was given to 406 tourists divided into 5 tourism sites at Chuncheon city, and logit model and probit model were used for analysis. The fitness levels of two models were very significant(p=0.0000). The study results suggest that the likelihood of the rural tourist to make a return visit is influenced by recognition of sustainable tourism, purchase of souvenir and farm produce, visitation of regional shops, conversation with regional residents, residents' participation on development, age and marriage. The results of such re-visitation demand can provide information for regional development strategies. The approach to re-visitation research impacts of consideration on sustainable tourism development is expected to become a useful foundation in studying on sustainable regional development.

• 전자통신동향분석
• /
• 제34권4호
• /
• pp.55-64
• /
• 2019

In this study, we introduce trends in neural-network-based deep learning research applied to dialogue systems. Recently, end-to-end trainable goal-oriented dialogue systems using long short-term memory, sequence-to-sequence models, among others, have been studied to overcome the difficulties of domain adaptation and error recognition and recovery in traditional pipeline goal-oriented dialogue systems. In addition, some research has been conducted on applying reinforcement learning to end-to-end trainable goal-oriented dialogue systems to learn dialogue strategies that do not appear in training corpora. Recent neural network models for end-to-end trainable chit-chat systems have been improved using dialogue context as well as personal and topic information to produce a more natural human conversation. Unlike previous studies that have applied different approaches to goal-oriented dialogue systems and chit-chat systems respectively, recent studies have attempted to apply end-to-end trainable approaches based on deep neural networks in common to them. Acquiring dialogue corpora for training is now necessary. Therefore, future research will focus on easily and cheaply acquiring dialogue corpora and training with small annotated dialogue corpora and/or large raw dialogues.

• 한국융합학회논문지
• /
• 제13권3호
• /
• pp.11-22
• /
• 2022

최근 교육 분야에서 다양한 인공지능 기술을 활용한 연구와 개발이 이뤄지고 있다. 인공지능을 활용한 교육 중 특히 대화형 에이전트는 시간과 공간의 제약을 받지 않고 음성인식, 번역과 같은 다양한 인공지능 기술과 결합해 더 효과적인 언어 학습을 가능하게 한다. 본 논문은 상용화된 교육용 플랫폼 중 이용자 수가 많고 영어 학습을 위한 대화형 에이전트가 활용된 플랫폼에 대한 동향 분석을 진행하였다. 동향 분석을 통해 현재 상용화된 교육용 플랫폼의 대화형 에이전트는 여러 한계점과 문제점이 존재했다. 구체적인 문제점과 한계점 분석을 위해 사전 학습된 최신 대용량 대화 모델과 비교 실험을 진행하였고, 실험 방법으로 대화형 에이전트의 대답이 사람과 비슷한지를 평가하는 Sensibleness and Specificity Average (SSA) 휴먼 평가를 진행하였다. 실험 내용을 바탕으로, 효과적인 학습을 위해 개선방안으로 대용량 파라미터로 학습된 대화 모델, 교육 데이터, 정보 검색 기능의 필요성을 제안했다.