• ETRI Journal
• /
• v.46 no.1
• /
• pp.137-153
• /
• 2024

This study introduces a new approach in Korean morphological analysis combining dictionary-based techniques with Transformer-based deep learning models. The key innovation is the use of a BERT-based reranking system, significantly enhancing the accuracy of traditional morphological analysis. The method generates multiple suboptimal paths, then employs BERT models for reranking, leveraging their advanced language comprehension. Results show remarkable performance improvements, with the first-stage reranking achieving over 20% improvement in error reduction rate compared with existing models. The second stage, using another BERT variant, further increases this improvement to over 30%. This indicates a significant leap in accuracy, validating the effectiveness of merging dictionary-based analysis with contemporary deep learning. The study suggests future exploration in refined integrations of dictionary and deep learning methods as well as using probabilistic models for enhanced morphological analysis. This hybrid approach sets a new benchmark in the field and offers insights for similar challenges in language processing applications.

• Steel and Composite Structures
• /
• v.51 no.2
• /
• pp.127-138
• /
• 2024

In this paper, structural behavior under fire conditions is comprehensively examined, and a novel software interface for testing interfaces efficiently is developed and validated. In order to accurately assess the response of structures to fire scenarios, advanced simulation techniques and modeling approaches are incorporated into the study. This interface enables accurate heat transfer analysis and thermo-mechanical simulations by integrating software tools such as CSI ETABS, CSI SAP2000, and OpenSees. Heat transfer models can be automatically generated, simulation outputs processed, and structural responses interpreted under a variety of fire scenarios using the proposed technique. As a result of rigorous testing and validation against established methods, including Cardington tests on scales and hybrid simulation approaches, the software interface has been proven to be effective and accurate. The analysis process is streamlined by this interface, providing engineers and researchers with a robust tool for assessing structural performance under fire conditions.

• Journal of Sensor Science and Technology
• /
• v.33 no.3
• /
• pp.119-124
• /
• 2024

Effective fire sensing is important to protect lives and property from the disaster. In this paper, we present an intelligent visual sensing method for detecting fires based on machine learning techniques. The proposed method involves a two-step process. In the first step, fire and non-fire images are used to train a convolutional neural network (CNN), and in the next step, feature vectors consisting of 256 values obtained from the CNN are used for the learning of a support vector machine (SVM). Linear and nonlinear SVMs with different parameters are intensively tested. We found that the proposed hybrid method using an SVM with a linear kernel effectively increased the recall rate of fire image detection without compromising detection accuracy when an imbalanced dataset was used for learning. This is a major contribution of this study because recall is important, particularly in the sensing of disaster situations such as fires. In our experiments, the proposed system exhibited an accuracy of 96.9% and a recall rate of 92.9% for test image data.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.40 no.4
• /
• pp.746-762
• /
• 2016

This study considers expression techniques and aesthetic values in the images of head dress reflected in a natural motif. The conclusions of the study are as follows. The first type is the realistic expression (52.4%) such as the reproduction of a natural object's essential form (27.0%), the partial derivation of the natural object (19.3%), and the planarization for the actual image of the natural object (6.1%). The second type is a metaphorical expression (39.0%) which emphasizes the morphological characteristics of nature (18.2%), the structuration of the natural object's silhouette into a three-dimensional or two-dimensional form (11.5%), and the abstract expression of the form in the natural object (9.3%). The third type is a hybrid expression (8.6%) that is a compromise between practical (or metaphorical expressions) so that expression techniques represent a compromise between the natural object's essential form and abstract expression (4.6%) or the combination of the natural object's silhouette into a three-dimensional or a two-dimensional visualization (4.0%). Aesthetic head dress values reflected in the natural motif first indicate a primitive value. This state of natural instinct recreates the natural object or combines part of the biological elements of the natural object to create an inducement to escape from the practical world. The second is amusement in the expression of animals in dynamic and humorous forms creates an illusion of animals being alive with a representative playful enjoyment. The third is abstraction that grant freedom in the observer's aesthetic rational through a reinterpretation of the fashion designer. The fourth is eclecticism where a compromise represents an act of mixing a variety of independent factors to create harmony with the imagery of nature created through the grafting of diverse expression techniques that break away from stereotypes of existing natural objects to create a type of nature that cultivates new values.

• International Journal of High-Rise Buildings
• /
• v.4 no.1
• /
• pp.9-25
• /
• 2015

Field measurements of various structures have been conducted for many purposes. Measurement data obtained by field measurement is very useful to determine vibration characteristics including dynamic characteristics such as the damping ratio, natural frequency, and mode shape of a structure. In addition, results of field measurements and modal identification can be used for modal updating of FEM analysis, for checking the efficiency of damping devices and so on. This paper shows some examples of field measurements and modal identification for structural health monitoring. As the first example, changes of dynamic characteristics of a 15-story office building in four construction stages from the foundation stage to completion are described. The dynamic characteristics of each construction stage were modeled as accurately as possible by FEM, and the stiffness of the main structural frame was evaluated and the FEM results were compared with measurements performed on non-load-bearing elements. Simple FEM modal updating was also applied. As the next example, full-scale measurements were also carried out on a high-rise chimney, and the efficiency of the tuned mass damper was investigated by using two kinds of modal identification techniques. Good correspondence was shown with vibration characteristics obtained by the 2DOF-RD technique and the Frequency Domain Decomposition method. As the last example, the wind-induced response using RTK-GPS and the feasibility of hybrid use of FEM analysis and RTK-GPS for confirming the integrity of structures during strong typhoons were shown. The member stresses obtained by hybrid use of FEM analysis and RTK-GPS were close to the member stresses measured by strain gauges.

• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.190-194
• /
• 2008

This paper presents a speaker identification technique which is one of the basic techniques of the ubiquitous robot companion. Though the conventional mel-frequency cepstral coefficients guarantee high performance of speaker identification in clean condition, the performance is degraded dramatically in noise condition. To overcome this problem, we employed the relative autocorrelation sequence mel-frequency cepstral coefficient which is one of the noise robust features. However, there are two problems in relative autocorrelation sequence mel-frequency cepstral coefficient: 1) the limited information problem. 2) the residual noise problem. In this paper, to deal with these drawbacks, we propose a multi-streaming method for the limited information problem and a hybrid method for the residual noise problem. To evaluate proposed methods, noisy speech is used in which air conditioner noise, classic music, and vacuum noise are artificially added. Through experiments, proposed methods provide better performance of speaker identification than the conventional methods.

• The Journal of the Korea Contents Association
• /
• v.8 no.7
• /
• pp.26-34
• /
• 2008

Due to spread of movies or online games which are applied with computer animation techniques, we can easily see scenes where numerous characters appear. In the case of large-scale crowd animation, if one were to increase reality of the scene, features of system would be lowered, and if one were to increase functioning of system, reality of the scene would be lowered. In realizing large-scale crowd animation with seafloor environment as background, the paper analyzed and applied elements that affect behavioral types of fishes; and by using concept of crowd, the paper enabled each group or object to control their behavioral type; by comparing and contrasting real-time calculation method as calculation method for animation and hybrid calculation method which is mixed calculation method, the paper seeks to find a method that increases functioning of the system while also expresses natural scenes.

• Journal of Korea Society of Industrial Information Systems
• /
• v.17 no.7
• /
• pp.111-118
• /
• 2012

Recently, most of firms have continuously released new products satisfying various needs of customers in order to increase market share. As a lot of products with various functionalities, prices and designs are released in the market, users have difficulties in choosing an appropriate product, especially for information technology driven devices. In case of digital cameras, inexperienced users spend a lot of time and efforts to find proper model for them. In this study, therefore, we design and implement a personalized recommendation system using analytic hierarchy process, one of the multi-criteria decision making techniques, and hybrid filtering combining content-based filtering and collaborative filtering to recommend a suitable product for inexperienced users of information technology devices.

• Journal of Chest Surgery
• /
• v.53 no.6
• /
• pp.346-352
• /
• 2020

Background: We aimed to assess the clinical outcomes of patients who underwent surgical repair of Kommerell diverticulum (KD) with individualized surgical methods. Methods: A retrospective analysis was performed of adult patients (aged ≥17 years) who underwent surgery to treat KD between June 2008 and October 2019. Results: Nine patients (median age, 45 years; range, 19-67 years; 7 men) underwent surgical repair. The indications for surgical therapy were acute aortic dissection in 2 patients, the presence of compressive symptoms due to dilated KD in 4 patients, and aneurysm growth in 3 patients. Various surgical techniques were used: (1) resection of the diverticulum stump and revascularization of the aberrant subclavian artery (n=3), (2) one-stage total-arch replacement including the diverticulum segment (n=3), and (3) hybrid repair (n=3). Early mortality occurred in 1 case of hybrid repair. Transient paraparesis occurred in a patient who underwent total arch repair as part of complicated acute aortic dissection. During follow-up (median duration, 30 months; range, 7-130 months), no late death or associated aortic complications were documented. All survivors were free from symptoms and had no abnormal findings on follow-up computed tomography. Conclusion: With a customized surgical approach and appropriate consideration of patient-specific anatomy and associated comorbidities, KD can be repaired with favorable outcomes.

• Journal of Computing Science and Engineering
• /
• v.7 no.4
• /
• pp.285-299
• /
• 2013

To enable hard real-time systems to take advantage of multicore processors, it is crucial to obtain the worst-case execution time (WCET) for programs running on multicore processors. However, this is challenging and complicated due to the inter-thread interferences from the shared resources in a multicore processor. Recent research used the combined cache conflict graph (CCCG) to model and compute the worst-case inter-thread interferences on a shared L2 cache in a multicore processor, which is called the CCCG-based approach in this paper. Although it can compute the WCET safely and accurately, its computational complexity is exponential and prohibitive for a large number of cores. In this paper, we propose three counter-based approaches to significantly reduce the complexity of the multicore WCET analysis, while achieving absolute safety with tightness close to the CCCG-based approach. The basic counter-based approach simply counts the worst-case number of cache line blocks mapped to a cache set of a shared L2 cache from all the concurrent threads, and compares it with the associativity of the cache set to compute the worst-case cache behavior. The enhanced counter-based approach uses techniques to enhance the accuracy of calculating the counters. The hybrid counter-based approach combines the enhanced counter-based approach and the CCCG-based approach to further improve the tightness of analysis without significantly increasing the complexity. Our experiments on a 4-core processor indicate that the enhanced counter-based approach overestimates the WCET by 14% on average compared to the CCCG-based approach, while its averaged running time is less than 1/380 that of the CCCG-based approach. The hybrid approach reduces the overestimation to only 2.65%, while its running time is less than 1/150 that of the CCCG-based approach on average.