• 한국언어정보학회:학술대회논문집
• /
• 한국언어정보학회 2007년도 정기학술대회
• /
• pp.192-201
• /
• 2007

This paper is a study on constructing a natural language interface to database, concentrating on generating textual answers. TGEN, a system that generates textual answer from query result tables is presented. The TGEN architecture guarantees its portability across domains. A combination of a frame-based approach and natural language generation techniques in the TGEN provides text fluency and text flexibility. The implementation result shows that this approach is feasible while a deep NLG approach is still far to be reached.

• 한국의상디자인학회지
• /
• 제22권3호
• /
• pp.49-61
• /
• 2020

This study analyzed the social media accounts and performed a Big Data analysis of Z-generation fashion using Textom Text Mining Techniques program and Ucinet Big Data analysis program. The research results are as follows: First, as a result of keyword analysis on 67.646 Z-generation fashion social media posts over the last 5 years, 220,211 keywords were extracted. Among them, 67 major keywords were selected based on the frequency of co-occurrence being greater than more than 250 times. As the top keywords appearing over 1000 times, were the most influential as the number of nodes connected to 'Z generation' (29595 times) are overwhelmingly, and was followed by 'millennials'(18536 times), 'fashion'(17836 times), and 'generation'(13055 times), 'brand'(8325 times) and 'trend'(7310 times) Second, as a result of the analysis of Network Degree Centrality between the key keywords for the Z-generation, the number of nodes connected to the "Z-generation" (29595 times) is overwhelmingly large. Next, many 'millennial'(18536 times), 'fashion'(17836 times), 'generation'(13055 times), 'brand'(8325 times), 'trend'(7310 times), etc. appear. These texts are considered to be important factors in exploring the reaction of social media to the Z-generation. Third, through the analysis of CONCOR, text with the structural equivalence between major keywords for Gen Z fashion was rearranged and clustered. In addition, four clusters were derived by grouping through network semantic network visualization. Group 1 is 54 texts, 'Diverse Characteristics of Z-Generation Fashion Consumers', Group 2 is 7 Texts, 'Z-Generation's teenagers Fashion Powers', Group 3 is 8 Texts, 'Z-Generation's Celebrity Fashions' Interest and Fashion', Group 4 named 'Gucci', the most popular luxury fashion of the Z-generation as one text.

• ETRI Journal
• /
• 제43권2호
• /
• pp.288-298
• /
• 2021

Abstractive text summarization is a process of making a summary of a given text by paraphrasing the facts of the text while keeping the meaning intact. The manmade summary generation process is laborious and time-consuming. We present here a summary generation model that is based on multilayered attentional peephole convolutional long short-term memory (MAPCoL; LSTM) in order to extract abstractive summaries of large text in an automated manner. We added the concept of attention in a peephole convolutional LSTM to improve the overall quality of a summary by giving weights to important parts of the source text during training. We evaluated the performance with regard to semantic coherence of our MAPCoL model over a popular dataset named CNN/Daily Mail, and found that MAPCoL outperformed other traditional LSTM-based models. We found improvements in the performance of MAPCoL in different internal settings when compared to state-of-the-art models of abstractive text summarization.

• 로봇학회논문지
• /
• 제16권3호
• /
• pp.276-282
• /
• 2021

The system for recognizing text in natural scenes has been applied in various industries. However, due to the change in brightness that occurs in nature such as light reflection and shadow, the text recognition performance significantly decreases. To solve this problem, we propose an adversarial shadow generation and training algorithm that is robust to shadow changes. The adversarial shadow generation and training algorithm divides the entire image into a total of 9 grids, and adjusts the brightness with 4 trainable parameters for each grid. Finally, training is conducted in a adversarial relationship between the text recognition model and the shaded image generator. As the training progresses, more and more difficult shaded grid combinations occur. When training with this curriculum-learning attitude, we not only showed a performance improvement of more than 3% in the ICDAR2015 public benchmark dataset, but also confirmed that the performance improved when applied to our's android application text recognition dataset.

• International Journal of Internet, Broadcasting and Communication
• /
• 제12권2호
• /
• pp.45-50
• /
• 2020

A Language model with neural networks commonly trained with likelihood loss. Such that the model can learn the sequence of human text. State-of-the-art results achieved in various language generation tasks, e.g., text summarization, dialogue response generation, and text generation, by utilizing the language model's next token output probabilities. Monotonous and boring outputs are a well-known problem of this model, yet only a few solutions proposed to address this problem. Several decoding techniques proposed to suppress repetitive tokens. Unlikelihood training approached this problem by penalizing candidate tokens probabilities if the tokens already seen in previous steps. While the method successfully showed a less repetitive generated token, the method has a large memory consumption because of the training need a big vocabulary size. We effectively reduced memory footprint by encoding words as sequences of subword units. Finally, we report competitive results with token level unlikelihood training in several automatic evaluations compared to the previous work.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권9호
• /
• pp.4568-4584
• /
• 2016

Steganography based on text generation has become a hot research topic in recent years. However, current text-generation methods which generate texts of normal style have either semantic or syntactic flaws. Note that texts of special genre, such as poem, have much simpler language model, less grammar rules, and lower demand for naturalness. Motivated by this observation, in this paper, we propose a text steganography that utilizes Markov chain model to generate Ci-poetry, a classic Chinese poem style. Since all Ci poems have fixed tone patterns, the generation process is to select proper words based on a chosen tone pattern. Markov chain model can obtain a state transfer matrix which simulates the language model of Ci-poetry by learning from a given corpus. To begin with an initial word, we can hide secret message when we use the state transfer matrix to choose a next word, and iterating until the end of the whole Ci poem. Extensive experiments are conducted and both machine and human evaluation results show that our method can generate Ci-poetry with higher naturalness than former researches and achieve competitive embedding rate.

• 대한음성학회지:말소리
• /
• 제53호
• /
• pp.75-92
• /
• 2005

Prosody Generation Based on C-ToBI Representation for Text-to-SpeechSeungwon Kim, Yu Zheng, Gary Geunbae Lee, Byeongchang KimProsody modeling is critical in developing text-to-speech (TTS) systems where speech synthesis is used to automatically generate natural speech. In this paper, we present a prosody generation architecture based on Chinese Tone and Break Index (C-ToBI) representation. ToBI is a multi-tier representation system based on linguistic knowledge to transcribe events in an utterance. The TTS system which adopts ToBI as an intermediate representation is known to exhibit higher flexibility, modularity and domain/task portability compared with the direct prosody generation TTS systems. However, the cost of corpus preparation is very expensive for practical-level performance because the ToBI labeled corpus has been manually constructed by many prosody experts and normally requires a large amount of data for accurate statistical prosody modeling. This paper proposes a new method which transcribes the C-ToBI labels automatically in Chinese speech. We model Chinese prosody generation as a classification problem and apply conditional Maximum Entropy (ME) classification to this problem. We empirically verify the usefulness of various natural language and phonology features to make well-integrated features for ME framework.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제31권12호
• /
• pp.1649-1659
• /
• 2004

텍스트 생성(text generation)은 언어가 아닌 다양한 지식원으로부터 텍스트를 생성해 내는 언어 처리의 한 분야로, 여러 가지 복합적이고 단계적인 과정을 통해 이루어진다. 본 논문에서는 자연스러운 텍스트 생성을 위한 여러 과정 중, 한번 언급된 대상(entity)을 자연스럽게 지시(refer)하기 위한 지시어 생성(referring expression generation), 특히 한국어에 두드러진 영형(zero pronoun)에 의한 대용화(pronominalization) 과정에 초점을 맞춘다. 이를 위해, 구문 정보와 비용기반 중심화 이론(cost-based centering model)을 바탕으로, 한국어에 적합한 지시어 특히 영형의 생성에 영향을 미치는 다양한 자질(feature)들온 규명하고, 기계 학습을 통해 지시어 생성 모델을 구축하였다. 세 개의 장르 - 묘사문(설명문), 뉴스, 짧은 우화 - 에서 총 95개의 텍스트로부터 학습이 이루어 졌으며 이론 대상으로, 제안된 자질들이 지시어의 생성, 특히 영형의 생성에 효율적으로 적용될 수 있음을 보였다. 또한, 지시어 생성과 관련된 기존의 방법론들과 본 논문에서 제안한 모델을 비교하여 성능이 크게 향상되었음을 보이고, T-test를 통해 99.9%의 신뢰 구간에서 그 성능 향상이 통계적으로 의미가 있음을 확인하였다.

• International Journal of Internet, Broadcasting and Communication
• /
• 제16권1호
• /
• pp.99-105
• /
• 2024

Runway released an updated version, Gen-2, in March 2023, which introduced new features that are different from Gen-1: it can convert text and images into videos, or convert text and images together into video images based on text instructions. This update will be officially open to the public in June 2023, so more people can enjoy and use their creativity. With this new feature, users can easily transform text and images into impressive video creations. However, as with all new technologies, comes the instability of AI, which also affects the results generated by Runway. This article verifies the feasibility of using Runway to generate the desired video from several aspects through personal practice. In practice, I discovered Runway generation problems and propose improvement methods to find ways to improve the accuracy of Runway generation. And found that although the instability of AI is a factor that needs attention, through careful adjustment and testing, users can still make full use of this feature and create stunning video works. This update marks the beginning of a more innovative and diverse future for the digital creative field.

• 실천공학교육논문지
• /
• 제15권2호
• /
• pp.509-517
• /
• 2023

TextRank 알고리즘을 활용한 연관 단어 추천 시스템과 사용자가 선택한 단어 기반 아이디어 생성 서비스를 반응형 웹으로 제공한다. 연관 단어 추천 시스템에서는 TextRank 알고리즘을 이용한 단어별 가중치 부여 방법 및 SoftMax를 적용한 확률 출력 방법을 논한다. 아이디어 생성 서비스에서는 mini-GPT를 이용한 아이디어 생성 방법과 인공지능 강화학습 방법에 대해 논한다. 반응형 웹에 대해서는 React와 Spring Boot, Flask 간의 연동 과정에 대해 논하며 전체적인 구동 방식에 대해 서술한다. 사용자가 원하는 주제를 입력하면 연관된 단어를 제공한다. 사용자는 연관된 단어를 선택하거나 원하는 단어를 추가하여 마인드맵을 구성한다. 사용자가 구성된 마인드맵에서 조합할 단어를 선택하면 새로 생성된 아이디어와 그와 연관된 특허를 제공한다. 본 웹서비스는 생성된 아이디어에 대해 다른 사용자와 공유할 수 있으며, 별점으로 사용자 피드백을 받아 인공지능을 개선한다.