• Journal of Information Management
• /
• v.40 no.4
• /
• pp.1-22
• /
• 2009

Large-scaled information extraction plays an important role in advanced information retrieval as well as question answering and summarization. Information extraction can be defined as a process of converting unstructured documents into formalized, tabular information, which consists of named-entity recognition, terminology extraction, coreference resolution and relation extraction. Since all the elementary technologies have been studied independently so far, it is not trivial to integrate all the necessary processes of information extraction due to the diversity of their input/output formation approaches and operating environments. As a result, it is difficult to handle scientific documents to extract both named-entities and technical terms at once. In this study, we define scientific as a set of 10 types of named entities and technical terminologies in a biomedical domain. in order to automatically extract these entities from scientific documents at once, we develop a framework for scientific core entity extraction which embraces all the pivotal language processors, named-entity recognizer, co-reference resolver and terminology extractor. Each module of the integrated system has been evaluated with various corpus as well as KEEC 2009. The system will be utilized for various information service areas such as information retrieval, question-answering(Q&A), document indexing, dictionary construction, and so on.

• Korean Journal of Food Science and Technology
• /
• v.47 no.3
• /
• pp.321-327
• /
• 2015

The optimal condition for preparation of powdered calcium acetate (LCCA) which has high solubility, from calcined powder (LCCP) of the littleneck clam shell by response surface methodology (RSM) was examined. Increased molar ratio of LCCP led to reduced solubility, yield, color values, and overall quality. The critical values of multiple response optimization of independent variables were 2.57 M of acetic acid and 1.57 M of LCCP. The actual values (pH 7.0, 96.1% for solubility, and 220.9% for yield) under the optimized condition were similar to the predicted values. LCCA showed strong buffering capacity between pH 4.89 and 4.92 on addition of ~2 mL of 1 N HCl. The calcium content and solubility of LCCA were 21.9-23.0 g/100 g and 96.1-100.1%, respectively. The FT-IR and XRD patterns of LCCA were identified as calcium acetate monohydrate, and FESEM images revealed an irregular and rod-like microstructure.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.13 no.1
• /
• pp.17-26
• /
• 2009

The capacity spectrum method (CSM), which is known to be an approximate technique for assessing the seismic capacity of an existing structure, was originally proposed for simple building structures that could be modeled as single-degree-of-freedom (SDOF) systems. More recently, however, CSM has increasingly been adopted for assessing most bridge structures, as it has many practical advantages. Some studies on this topic are now being performed, and a few results of these have been presented as ground-breaking research. However, studies have until now been limited to symmetrical straight bridges only. This study evaluates the practical applicability of CSM to the evaluation of irregular curved bridges. For this purpose, the seismic capacities of 3-span prestressed concrete bridges with different subtended angles subjected to some recorded earthquakes are compared with a more refined approach based on nonlinear time history analysis. The results of the study show that when used for curved bridges, CSM induces higher inelastic displacement responses than the actual values, and that the gap between the two becomes larger as the subtended angle increases.

• Journal of the Korea Concrete Institute
• /
• v.24 no.2
• /
• pp.147-156
• /
• 2012

In recent years, many high-rise buildings have been constructed in irregular structural system with inclined columns, which may have effect on the structural behavior of beam-column joints. Since the external load leads to shear and flexural forces on the inclined columns in different way from those on the conventional vertical columns, failure mode, resistant strength, and ductility capacity of the inclined column-beam joints may be different than those of the perpendicular beam-column joints. In this study, six RC inclined beam-column joint specimens were tested. The main parameter of the specimens was the angle between axes of the column and beam (90, 67.5, and 45 degree). Test results indicated that the structural behavior of conventional perpendicular beam-column joint was different to that of the inclined beam-column joints, due to different loading conditions between inclined and perpendicular beam-column joints. Both upper and lower columns of perpendicular beam-column joints were subjected to compressive force, while the upper and lower columns of the inclined beam-column joints were subjected to tensile and compressive forces, respectively.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.44 no.6
• /
• pp.888-895
• /
• 2015

For effective utilization of butter clam shell as a natural calcium resource, the optimal conditions for preparation of calcium acetate (BCCA) with high solubility were determined using response surface methodology (RSM). The polynomial models developed by RSM for pH, solubility, and yield were highly effective in describing the relationships between factors (P<0.05). Increased molar ratio of calcined powder (BCCP) from butter clam shell led to reduction of solubility, yield, color values, and overall quality. Critical values of multiple response optimization to independent variables were 2.70 M and 1.05 M for acetic acid and BCCP, respectively. The actual values (pH 7.04, 93.0% for solubility and 267.5% for yield) under optimization conditions were similar to predicted values. White indices of BCCAs were in the range of 89.7~93.3. Therefore, color value was improved by calcination and organic acid treatment. Buffering capacity of BCCAs was strong at pH 4.88 to 4.92 upon addition of ~2 mL of 1 N HCl. Calcium content and solubility of BCCAs were 20.7~22.8 g/100 g and 97.2~99.6%, respectively. The patterns of fourier transform infrared spectrometer and X-ray diffractometer analyses from BCCA were identified as calcium acetate monohydrate, and microstructure by field emission scanning electron microscope showed an irregular form.

• Korean Journal of Construction Engineering and Management
• /
• v.22 no.5
• /
• pp.73-85
• /
• 2021

This study proposes an efficient management direction for Korean construction accident cases through a deep learning-based text data classification model. A deep learning model was developed, which categorizes five categories of construction accidents: fall, electric shock, flying object, collapse, and narrowness, which are representative accident types of KOSHA. After initial model tests, the classification accuracy of fall disasters was relatively high, while other types were classified as fall disasters. Through these results, it was analyzed that 1) specific accident-causing behavior, 2) similar sentence structure, and 3) complex accidents corresponding to multiple types affect the results. Two accuracy improvement experiments were then conducted: 1) reclassification, 2) elimination. As a result, the classification performance improved with 185.7% when eliminating complex accidents. Through this, the multicollinearity of complex accidents, including the contents of multiple accident types, was resolved. In conclusion, this study suggests the necessity to independently manage complex accidents while preparing a system to describe the situation of future accidents in detail.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.15 no.1
• /
• pp.56-61
• /
• 2022

In this paper, we classify and present AI algorithms and natural language processing methods used in chatbots. A framework that can be used to implement a chatbot is also described. A chatbot is a system with a structure that interprets the input string by constructing the user interface in a conversational manner and selects an appropriate answer to the input string from the learned data and outputs it. However, training is required to generate an appropriate set of answers to a question and hardware with considerable computational power is required. Therefore, there is a limit to the practice of not only developing companies but also students learning AI development. Currently, chatbots are replacing the existing traditional tasks, and a practice course to understand and implement the system is required. RNN and Char-CNN are used to increase the accuracy of answering questions by learning unstructured data by applying technologies such as deep learning beyond the level of responding only to standardized data. In order to implement a chatbot, it is necessary to understand such a theory. In addition, the students presented examples of implementation of the entire system by utilizing the methods that can be used for coding education and the platform where existing developers and students can implement chatbots.

• Design Convergence Study
• /
• v.16 no.6
• /
• pp.169-184
• /
• 2017

There is a lot of needs that are not expressed as much as the expressed needs in familiar products and services that are used in daily life such as a smartphone. Finding the 'Inconveniences in familiar use' make it possible to create opportunities for value expanding in the existing products and service area. There are a lot of related works, which have studied the definition of hidden needs and the methods to find it. But, they are making it difficult to address the hidden needs in the cases of familiar use due to focus on the new product or service developing typically. In this study, we try to redefine the hidden needs in the daily familiarity and approach it in the new way to find out. Because of the users' unability to express what they want and the complexity of needs which can not be explained clearly, we can not approach it as the quantitative issue. For this reason, the basic data type selected as the user behavior data excluding all description is the screen-shot of the smartphone. We try to apply the integrated rules and patterns to the individual data using the qualitative coding techniques to overcome the limitations of qualitative analysis based on unstructured data. From this process, We can not only extract meaningful clues which can make to understand the hidden needs but also identify the possibility as a way to discover hidden needs through the review of relevance to actual market trends. The process of finding hidden needs is not easy to systemize in itself, but we expect the possibility to be conducted a reference frame for finding hidden needs of other further studies.

• Journal of Intelligence and Information Systems
• /
• v.29 no.4
• /
• pp.165-184
• /
• 2023

Recently, Deep learning analysis of unstructured text data using language models, such as Google's BERT and OpenAI's GPT has shown remarkable results in various applications. Most language models are used to learn generalized linguistic information from pre-training data and then update their weights for downstream tasks through a fine-tuning process. However, some concerns have been raised that privacy may be violated in the process of using these language models, i.e., data privacy may be violated when data owner provides large amounts of data to the model owner to perform fine-tuning of the language model. Conversely, when the model owner discloses the entire model to the data owner, the structure and weights of the model are disclosed, which may violate the privacy of the model. The concept of offsite tuning has been recently proposed to perform fine-tuning of language models while protecting privacy in such situations. But the study has a limitation that it does not provide a concrete way to apply the proposed methodology to text classification models. In this study, we propose a concrete method to apply offsite tuning with an additional classifier to protect the privacy of the model and data when performing multi-classification fine-tuning on Korean documents. To evaluate the performance of the proposed methodology, we conducted experiments on about 200,000 Korean documents from five major fields, ICT, electrical, electronic, mechanical, and medical, provided by AIHub, and found that the proposed plug-in model outperforms the zero-shot model and the offsite model in terms of classification accuracy.

• Korean Chemical Engineering Research
• /
• v.49 no.1
• /
• pp.75-82
• /
• 2011

Five different composition UV-cured poly(urethane acrylate-co-acrylic acid) (PU-co-AA) films have been prepared by reacting isophorone diisocyanate(IPDI), polycaprolactone triol(PCLT), 2-hydroxyethyl acrylate(HEA), and different weight ratio trimethylolpropane triacrylate(TMPTA) and acrylic acid(AA) as diluents, and characterized using a Fourier transform infrared spectroscopy(FT-IR). The adhesion properties onto the stainless steel, morphology, mechanical hardness, and electrical property of UV-cured PU-co-AA films were investigated as a function of acrylic acid(AA) content. All the PU-co-AA films are structure-less and the molecular ordering and packing density decreased with increasing content of AA due to the flexible structure and -COOH side chains in AA. The crosscut test showed that PU-co-AA films without AA and with low content of AA showed 0% adhesion(0B) and the adhesion of PU-co-AA films in the range of 40-50% AA increased dramatically as the content of AA increases. The pull-off measurements showed that the adhesion force of PU-co-AA films to stainless steel substrate varied from 6 to 31 kgf /$cm^2$ and increased linearly with increasing AA content. The mechanical hardness also decreased as the content of AA increases. This may come from relatively linear and flexible structure in AA and low crystallinity in PU-co-AA films with higher content of AA. The higher AA-containing PU-co-AA films showed higher dielectric constant due to the increase of polarization by introducing AA monomer. In conclusion, the physical properties of UV-cured PU-co-AA films are strongly dependent upon the content of AA and the incorporation of AA in polyurethane acrylate is very useful way to increase the adhesion strength of UV-curable polymers on the stainless steel substrate.