• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.41 no.3
• /
• pp.67-77
• /
• 2004

The recent research has been focused on fusion of audio and visual features for reliable speech recognition in noisy environments. In this paper, we propose a neural network based model of robust speech recognition by integrating audio, visual, and contextual information. Bimodal Neural Network(BMNN) is a multi-layer perception of 4 layers, each of which performs a certain level of abstraction of input features. In BMNN the third layer combines audio md visual features of speech to compensate loss of audio information caused by noise. In order to improve the accuracy of speech recognition in noisy environments, we also propose a post-processing based on contextual information which are sequential patterns of words spoken by a user. Our experimental results show that our model outperforms any single mode models. Particularly, when we use the contextual information, we can obtain over 90% recognition accuracy even in noisy environments, which is a significant improvement compared with the state of art in speech recognition. Our research demonstrates that diverse sources of information need to be integrated to improve the accuracy of speech recognition particularly in noisy environments.

• Speech Sciences
• /
• v.10 no.3
• /
• pp.37-45
• /
• 2003

This study attempted to determine the effects of a context-based drawing task on the language expression of a severe Wernicke's and Broca's aphasic. The subjects in this study showed a poor auditory comprehension and naming performance. They also showed paraphasia and perseveration. This study focused on improving language expression by a drawing task based on conversation at hand. Ten target words were chosen which were easily drawnable and familiar to the subjects. The results showed that the context-based drawing task was effective on improving the subjects' confrontation naming ability and expressive language ability in terms of explanation of sentences. In addition, the Broca's aphasic showed improved naming ability when the contextual cues were given and he was supposed to spontaneously name words.

• Korean Journal of Cognitive Science
• /
• v.24 no.2
• /
• pp.133-171
• /
• 2013

The gist of a scene can be identified in less than 100msec, and violation in the gist can influence the way to allocate attention to the parts of a scene. In other words, people tend to allocate more attention to the object(s) inconsistent with the gist of a scene and to have better memory of them. To investigate the effects of contextual consistency on the attention allocation and object memory, two experiments were conducted. In both experiments, a $3{\times}2$ factorial design was used with scene presentation time(2s, 5s, and 10s) as a between-subject factor and object-background contextual consistency(consistent, inconsistent) as a within-subject factor. In Experiment 1, eye movements were recorded while the participants viewed line-drawing scenes. The results showed that the eye movement patterns were different according to whether the scenes were consistent or not. Context-inconsistent objects showed faster initial fixation indices, longer fixation times, more frequent returns than context-consistent ones. These results are entirely consistent with those of previous studies. If an object is identified as inconsistent with the gist of a scene, it attracts attention. Furthermore, the inconsistent objects and their locations in the scenes were recalled better than the consistent ones and their locations. Experiment 2 was the same as Experiment 1 except that a dual-task paradigm was used to reduce the amount of attention to allocate to the objects. Participants had to detect the positions of the probe occurring every second while they viewed the scenes. Nonetheless, the result patterns were the same as in Experiment 1. Even when the amount of attention to allocate to the scene contents was reduced, the same effects of contextual inconsistency were observed. These results indicate that the object-background contextual consistency has a strong influence on the way of allocating attention and the memory of objects in a scene.

• Journal of Information Technology Services
• /
• v.8 no.1
• /
• pp.203-215
• /
• 2009

This paper illustrates the application of co-occurrence theory to generate lightweight ontologies semi-automatically. The proposed model includes three steps of a (Semi-) Automatic creation of Ontology; (they are conceptually named as) the Syntactic-based Ontology, the Semantic-based Ontology and the Ontology Refinement. Each of these three steps are designed to interactively work together, so as to generate Lightweight Ontologies. The Syntactic-based Ontology step includes generating Association words using co-occurrence in web documents. The Semantic-based Ontology step includes the Alignment large Association words with small Ontology, through the process of semantic relations by contextual terms. Finally, the Ontology Refinement step includes the domain expert to refine the lightweight Ontologies. We also conducted a case study to generate lightweight ontologies in specific domains(news domain). In this paper, we found two directions including (1) employment co-occurrence theory to generate Syntactic-based Ontology automatically and (2) Alignment large Association words with small Ontology to generate lightweight ontologies semi-automatically. So far as the design and the generation of big Ontology is concerned, the proposed research will offer useful implications to the researchers and practitioners so as to improve the research level to the commercial use.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.14 no.2
• /
• pp.919-926
• /
• 2013

The relation of City and type caused by Neo-Rationalism often posed a subject in the modern urban architectural design and is treated by essential concept and methodology. The purpose of this study was that Neo-Rationalism architecture recognized and treated by metaphysical form and singular image examined on the urban viewpoint. In other words, interpretation concept of city and architecture considered through typological architectural theory and work analysis of Neo-Rationalism architect. Accordingly, the urban contextual characteristics of Neo-Rationalism architecture were aimed to analyze by examining the acceptance mode.

• Journal of KIISE
• /
• v.42 no.11
• /
• pp.1452-1458
• /
• 2015

Korean Morphological Analysis is a difficult process. Because Korean is an agglutinative language, one of the most important processes in Morphological Analysis is Morpheme Recovery. There are some methods using Heuristic rules and Pre-Analyzed Partial Words that were examined for this process. These methods have performance limits as a result of not using contextual information. In this study, we built a Korean morpheme recovery system using deep learning, and this system used word embedding for the utilization of contextual information. In '들/VV' and '듣/VV' morpheme recovery, the system showed 97.97% accuracy, a better performance than with SVM(Support Vector Machine) which showed 96.22% accuracy.

• Journal of Internet Computing and Services
• /
• v.21 no.3
• /
• pp.123-131
• /
• 2020

Along with the rapid development of artificial intelligence technology, natural language processing, which deals with human language, is also actively studied. In particular, BERT, a language model recently proposed by Google, has been performing well in many areas of natural language processing by providing pre-trained model using a large number of corpus. Although BERT supports multilingual model, we should use the pre-trained model using large amounts of Korean corpus because there are limitations when we apply the original pre-trained BERT model directly to Korean. Also, text contains not only vocabulary, grammar, but contextual meanings such as the relation between the front and the rear, and situation. In the existing natural language processing field, research has been conducted mainly on vocabulary or grammatical meaning. Accurate identification of contextual information embedded in text plays an important role in understanding context. Knowledge graphs, which are linked using the relationship of words, have the advantage of being able to learn context easily from computer. In this paper, we propose a system to extract Korean contextual information using pre-trained BERT model with Korean language corpus and knowledge graph. We build models that can extract person, relationship, emotion, space, and time information that is important in the text and validate the proposed system through experiments.

• Korean Journal of Cognitive Science
• /
• v.27 no.1
• /
• pp.129-157
• /
• 2016

This study investigates speech perception in noise (SPIN) in Korean. A new type of Korean SPIN test was developed by adopting a similar format to the English SPIN test. The predictability effects, noise effects and their interactions were examined in order to verify the previous findings based on English. The data from 14 Korean adults collected with this new type of Korean SPIN test confirmed the previous findings: first, the participants' overall performance was better in low noise conditions than in high noise conditions. Secondly, there was a tendency for highly predictable words to be more accurately perceived than less predictable words especially in high noise conditions. The results were interpreted in such a way that the listeners actively used both types of information: acoustic information and contextual information in speech perception. When the acoustic property of the speech sound was degraded with noise, the listeners took advantage of the linguistic contextual information in their processing of the speech sound. The findings of this study conform to those of the previous studies based on the English SPIN test. In addition, a possible effect of the frequency of target word was also found, calling for further investigation in this field of research in Korean. Implications of the results were also discussed. (Cyber Hankuk University of Foreign Studies)

• Journal of Internet Computing and Services
• /
• v.12 no.5
• /
• pp.171-179
• /
• 2011

In this paper it is hypothesized that word stimuli that are presented by Google’s search word, would improve the quality of the design solution, so this research examines the effect of related search word stimuli in concept generation and analyzes the results through the processes of creative cognition. In the process of concept generation, words are given as stimuli which are generated through Google's related search and these search words are given by 5 levels. Google search is based on the collaboration philosophy. People's participation and contribution recreate knowledge and information, so these renewed and related search words update in real time by people are used as stimuli. Two problems are provided with related search words. After the design concept generation the results are analyzed by 3 bases: the usage of related search words and those of frequency, creativity, and Finke's 12 Geneplore model. These are the results of the research. Many levels of related search words are used in design concept generation but especially higher levels which are more related to search words are more used than lower levels. The usage of multi words and conjunction with higher levels and lower levels words are observed in creative results. On the creative cognitive processes, it is more creative when using association and mental transformation with the related search words than using the related search words simply. Creative outputs also use conceptual interpretation, functional inference, and contextual shifting of creative cognitive processes of Finke's 12 Geneplore model.

• The Transactions of the Korea Information Processing Society
• /
• v.5 no.10
• /
• pp.2591-2599
• /
• 1998

In this paper, we proose a method of recognizing unknown words and correcting spelling errors(including spacing erors) to increase the performance of Korean information processing systems. Unknown words are recognized through comparative analysis of two or more morphologically similar eojeols(spacing units in Korean) including the same unknown word candidates. And spacing errors and spelling errors are corrected by using lexicatlized rules shich are automatically extracted from very large raw corpus. The extractionof the lexicalized rules is based on morphological and contextual similarities between error eojeols and their corection eojeols which are confirmed to be used in the corpus. The experimental result shows that our system can recognize unknown words in an accuracy of 98.9%, and can correct spacing errors and spelling errors in accuracies of 98.1% and 97.1%, respectively.