• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.931-934
• /
• 2019

Research using artificial intelligence to generate captions for an image has been studied extensively. However, these systems are unable to create creative stories that include more than one sentence based on image content. A story is a better way that humans use to foster social cooperation and develop social norms. This paper proposes a framework that can generate a relatively short story to describe based on the context of an image. The main contributions of this paper are (1) An unsupervised framework which uses recurrent neural network structure and encoder-decoder model to construct a short story for an image. (2) A huge English novel dataset, including horror and romantic themes that are manually collected and validated. By investigating the short stories, the proposed model proves that it can generate more creative contents compared to existing intelligent systems which can produce only one concise sentence. Therefore, the framework demonstrated in this work will trigger the research of a more robust AI story writer and encourages the application of the proposed model in helping story writer find a new idea.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.2
• /
• pp.93-100
• /
• 2022

Keyword-oriented search methods are mainly used as data search methods, but this is not suitable as a search method in the legal field where professional terms are widely used. In response, this paper proposes an effective data search method in the legal field. We describe embedding methods optimized for determining similarities between sentences in the field of natural language processing of legal domains. After embedding legal sentences based on keywords using TF-IDF or semantic embedding using Universal Sentence Encoder, we propose an optimal way to search for data by combining BERT models to check similarities between sentences in the legal field.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.389-392
• /
• 2021

문장 압축은 원본 문장의 중요한 의미를 보존하는 짧은 길이의 압축 문장을 생성하는 자연어처리 태스크이다. 문장 압축은 사용자가 텍스트로부터 필요한 정보를 빠르게 획득할 수 있도록 도울 수 있어 활발히 연구되고 있지만, 기존 연구들은 사람이 직접 정의한 압축 규칙이 필요하거나, 모델 학습을 위해 대량의 데이터셋이 필요하다는 문제점이 존재한다. 사전 학습된 언어 모델을 통한 perplexity 기반의 문장 점수 측정을 통해 문장을 압축하여 압축 규칙과 모델 학습을 위한 데이터셋이 필요하지 않은 연구 또한 존재하지만, 문장 점수 측정에 문장에 속한 단어들의 의미적 중요도를 반영하지 못하여 중요한 단어가 삭제되는 문제점이 존재한다. 본 논문은 언어 정보 중 품사 정보, 의존관계 정보, 개체명 정보의 중요도를 수치화하여 perplexity 기반의 문장 점수 측정에 반영하는 방법을 제안한다. 또한 제안한 문장 점수 측정 방법을 활용하였을 때 문장 점수 측정 기반 문장 압축 모델의 문장 압축 성능이 향상됨을 확인하였으며, 이를 통해 문장에 속한 단어의 언어 정보를 문장 점수 측정에 반영하는 것이 의미적으로 적절한 압축 문장을 생성하는 데 도움이 될 수 있음을 보였다.

• Journal of KIISE
• /
• v.42 no.12
• /
• pp.1568-1574
• /
• 2015

Relation extraction plays a role in for the process of transforming a sentence into a form of knowledge base. In this paper, we focus on predicates in a sentence and aim to identify the relevant knowledge base properties required to elucidate the relationship between entities, which enables a computer to understand the meaning of a sentence more clearly. Distant Supervision is a well-known approach for relation extraction, and it performs lexicalization tasks for knowledge base properties by generating a large amount of labeled data automatically. In other words, the predicate in a sentence will be linked or mapped to the possible properties which are defined by some ontologies in the knowledge base. This lexical and ontological linking of information provides us with a way of generating structured information and a basis for enrichment of the knowledge base.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.6
• /
• pp.761-766
• /
• 2011

A hedge is a linguistic device to express uncertainties. Hedges are used in a sentence when the writer is uncertain or has doubt about the contents of the sentence. Due to this uncertainty, sentences with hedges are considered to be non-factual. There are many applications which need to determine whether a sentence is factual or not. Detecting hedges has the advantage in information retrieval, and information extraction, and QnA systems, which make use of non-hedge sentences as target to get more accurate results. In this paper, we constructed Korean hedge corpus, and extracted generalized hedge cue-word patterns from the corpus, and then used them in detecting hedges. In our experiments, we achieved 78.6% in F1-measure.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.1
• /
• pp.52-57
• /
• 2022

Speech consists of the sum of complex sine-waves. This study investigated the perception of electrical hearing by applying the sinusoidal model to cochlear implant simulation. Fourteen adults with normal hearing participated in this study. The sentence recognition tests were implemented using the sentence lists processed by the sinusoidal model which extracts 2, 4, 6, 8 sine-wave components and sentence lists processed by the same sinusoidal model along with cochlear implant simulation (8 channel vocoders). The results showed lower speech recognition for the sentence lists processed by the sinusoidal model and cochlear implant simulation compared to those by the sinusoidal model alone. Notably, the lower the number of sine-wave components (2), the larger the difference was. This study provides the perceptual pattern of sine-wave speech for electrical hearing by cochlear implant listeners, and basic data for development of speech processing algorithms in cochlear implants.

• Korean Journal of Cognitive Science
• /
• v.3 no.1
• /
• pp.41-60
• /
• 1991

Most of the PC users manipulate the computer by using a few commands which are familiar with them. However by using Hangul sentences instead of using DOS commands, the optimal commands can be generated and flexibility can be provided. For this purpose, the conversion method of the input sentence into DOS commands is studied by means of morphological analysis, syntactic analysis, semantic analysis, and conceptual analysis. Tabular parsing is used in morphological analysis. case grammar is used in syntactic and semantic analysis. Case grammar is used in syntactic and semantic analysis. The meaning of sentence is represeented by the semantic network, from which we can generate a sequence DOS commands.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2007.11a
• /
• pp.249-258
• /
• 2007

This paper presents an implementation of a gramar of Dynamic Syntax for Japanese. Dynamic Syntax is a grammar formalism which enables a parser to process a sentence in an incremental fashion, establishing the semantic representation. Currently the application of lexical rules and transition rules in Dynamic Syntax is carried out arbitrarily and this leads to inefficient parsing. This paper provides an algorithm of rule application and partitioned parsing state for efficient parsing with special reference to processing Japanese, which is one of head-final languages. At the present stage the parser is still small but can parse scrambled sentences, relative clause constructions, and embedded clauses. The parser is written in Prolog and this paper shows that the parser can process null arguments in a complex sentence in Japanese.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2007.11a
• /
• pp.285-292
• /
• 2007

The parallel corpus is an important resource in the research field of data-driven natural language processing, but there are only a few parallel corpora publicly available nowadays, mostly due to the high labor force needed to construct this kind of resource. A novel strategy is brought out to automatically fetch parallel text from the web in this paper, which may help to solve the problem of the lack of parallel corpora with high quality. The system we develop first downloads the web pages from certain hosts. Then candidate parallel page pairs are prepared from the page set based on the outer features of the web pages. The candidate page pairs are evaluated in the last step in which the sentences in the candidate web page pairs are extracted and aligned first, and then the similarity of the two web pages is evaluate based on the similarities of the aligned sentences. The experiments towards a multilingual web site show the satisfactory performance of the system.

• Phonetics and Speech Sciences
• /
• v.7 no.3
• /
• pp.55-64
• /
• 2015

Previous research on syntactic processing shows that the interpretation of a syntactically ambiguous sentence is frequently strongly biased towards one meaning over another. The current study investigated the effect of bias strength on the production of disambiguating prosody for English ambiguous sentences. In Experiment 1, 40 speakers gave default readings of 18 syntactically ambiguous sentences. Questioning was used to prove intended meanings behind default readings. Intended meanings were treated as interpretation biases when a majority of speakers read a sentence with the same intended meaning. The size of the majority was used to establish bias strength. In Experiment 2, 10 speakers were instructed to use prosody to disambiguate given alternate meanings of the sentences from Experiment 1. The results indicated an effect of bias strength on disambiguating prosody: speakers used temporal juncture cues to reliably disambiguate alternate meanings for sentences with a weak interpretation bias, but not for those with a strong bias. Overall, the results indicated that interpretation biases strongly affect the production of prosody.