• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.463-467
• /
• 2023

사전 학습 모델을 특정 데이터에 미세 조정할 때, 최대 길이는 사전 학습에 사용한 최대 길이 파라미터를 그대로 사용해야 한다. 이는 상대적으로 긴 시퀀스의 처리를 요구하는 일부 작업에서 단점으로 작용한다. 본 연구는 상대적으로 긴 시퀀스의 처리를 요구하는 질의 응답(Question Answering, QA) 작업에서 사전 학습 모델을 활용할 때 발생하는 시퀀스 길이 제한에 따른 성능 저하 문제를 극복하는 방법론을 제시한다. KorQuAD v1.0과 AIHub에서 확보한 데이터셋 4종에 대하여 BERT와 RoBERTa를 이용해 성능을 검증하였으며, 실험 결과, 평균적으로 길이가 긴 문서를 보유한 데이터에 대해 성능이 향상됨을 확인할 수 있었다.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.1
• /
• pp.63-72
• /
• 2013

This study it to verify the validity of Korean version of DELES (distance education learning environment survey) and analyze its relationship with learning achievement and distance education satisfaction. The target population of this study is students of K cyber university and a total of 254 cases are used for the analysis. Exploratory and confirmatory factor analysis is applied to verify 6 factors of DELES and structural equation analysis is applied to examine the relationship between distance education learning environment and learning achievement and distance education satisfaction. The study result shows that DELES is composed of six factors such as instructor support, student interaction & collaboration, personal relevance, authentic learning, active learning and student autonomy and its model fits are appropriate. The result of structural equation analysis shows distance education learning environment significantly influences distance education satisfaction directly as well as indirectly mediated by learning achievement. Learning achievement also significantly influences distance education satisfaction. Conclusions and implications are followed.

• Korean Journal of Remote Sensing
• /
• v.37 no.5_1
• /
• pp.871-884
• /
• 2021

The purpose of this study was to determine ways to increase efficiency in constructing and verifying artificial intelligence learning data on land cover using aerial and satellite images, and in applying the data to AI learning algorithms. To this end, multi-resolution datasets of 0.51 m and 10 m each for 8 categories of land cover were constructed using high-resolution aerial images and satellite images obtained from Sentinel-2 satellites. Furthermore, fine data (a total of 17,000 pieces) and coarse data (a total of 33,000 pieces) were simultaneously constructed to achieve the following two goals: precise detection of land cover changes and the establishment of large-scale learning datasets. To secure the accuracy of the learning data, the verification was performed in three steps, which included data refining, annotation, and sampling. The learning data that wasfinally verified was applied to the semantic segmentation algorithms U-Net and DeeplabV3+, and the results were analyzed. Based on the analysis, the average accuracy for land cover based on aerial imagery was 77.8% for U-Net and 76.3% for Deeplab V3+, while for land cover based on satellite imagery it was 91.4% for U-Net and 85.8% for Deeplab V3+. The artificial intelligence learning datasets on land cover constructed using high-resolution aerial and satellite images in this study can be used as reference data to help classify land cover and identify relevant changes. Therefore, it is expected that this study's findings can be used in the future in various fields of artificial intelligence studying land cover in constructing an artificial intelligence learning dataset on land cover of the whole of Korea.

• Journal of The Korean Association of Information Education
• /
• v.25 no.3
• /
• pp.459-469
• /
• 2021

In this paper, a case study of basic data science practice education for field teachers and pre-service teachers was studied. In this paper, for basic data science education, spreadsheet software was used as a data collection and analysis tool. After that, we trained on statistics for data processing, predictive hypothesis, and predictive model verification. In addition, an educational case for collecting and processing thousands of public big data and verifying the population prediction hypothesis and prediction model was proposed. A 34-hour, 17-week curriculum using a spreadsheet tool was presented with the contents of such basic education in data science. As a tool for data collection, processing, and analysis, unlike Python, spreadsheets do not have the burden of learning program- ming languages and data structures, and have the advantage of visually learning theories of processing and anal- ysis of qualitative and quantitative data. As a result of this educational case study, three predictive hypothesis test cases were presented and analyzed. First, quantitative public data were collected to verify the hypothesis of predicting the difference in the mean value for each group of the population. Second, by collecting qualitative public data, the hypothesis of predicting the association within the qualitative data of the population was verified. Third, by collecting quantitative public data, the regression prediction model was verified according to the hypothesis of correlation prediction within the quantitative data of the population. And through the satisfaction analysis of pre-service and field teachers, the effectiveness of this education case in data science education was analyzed.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.05a
• /
• pp.315-318
• /
• 2006

본 논문에서는 TV OSD(On Screen Display) 메뉴 자동검증 시스템에서 문자패턴의 실시간 인식을 위한 방법론을 고찰한다. 이는 일반적인 문자인식 문제와는 달리 시스템 환경에 대한 몇 가지 가정과 제약조건을 고려해야 한다. 예컨대 문제의 특성상 카메라 및 TV 제어 기기부의 동작과 연동하는 작업 스케쥴링 기능과 실시간 분석기능 등의 요건은 시스템개발을 복잡하게 하는 반면, 주어진 OSD 메뉴 데이터로부터 검증과정은 미지 패턴에 대한 인식과정을 단순화하여 일종의 판정(decision) 문제로 고려될 수 있게 한다. 본 연구에서는 인식의 방법론으로서 수정된 구조의 FMM 신경망을 적용한다. 이는 하이퍼박스 기반의 패턴 분류기로서 간결하면서도 강력한 학습기능을 제공한다. 기존의 FMM 모델이 갖는 단점인 학습패턴에서 특징분포와 빈도를 고려하지 못한다는 점을 개선하여, 특징과 하이퍼박스간의 가중치 요소를 고려한 활성화 특성을 정의한다. 또한 실제 데이터를 사용한 실험결과를 통해 제안된 이론의 유용성을 고찰한다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.1336-1338
• /
• 2022

폐기물 관리는 전 세계적으로 환경, 사회, 경제 문제를 일으키고 있다. 이러한 문제를 예방하고자 폐기물을 효율적으로 관리하기 위해, 인공지능을 통한 연구를 제안하고 있다. 따라서 본 논문에서는 GAN 기반 데이터 증강을 통한 폐기물 객체 인식모델을 제안한다. Open Images Dataset V6와 AI Hub의 공공 데이터 셋을 융합하여 폐기물 품목에 해당하는 이미지들을 정제하고 라벨링한다. 이때, 실제 배출환경에서 발생할 수 있는 장애물로 인한 일부분만 노출된 폐기물, 부분 파손, 눕혀져 배출, 다양한 색상 등의 인식저해요소를 모델 학습에 반영할 수 있도록 일반적인 데이터 증강과 GAN을 통한 데이터 증강을 병합 사용한다. 이후 YOLOv4 기반 폐기물 이미지 인식 모델 학습을 진행하고, 학습된 이미지 인식 모델에 대한 검증 및 평가를 mAP, F1-Score로 진행한다. 이를 통해 향후 스마트폰 애플리케이션과 융합하여 효율적인 폐기물 관리 체계를 구축할 수 있을 것이다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.5
• /
• pp.533-539
• /
• 2019

This study proposes a malware classification model that can handle arbitrary length input data using the Microsoft Malware Classification Challenge dataset. We are based on imaging existing data from malware. The proposed model generates a lot of images when malware data is large, and generates a small image of small data. The generated image is learned as time series data by Dynamic RNN. The output value of the RNN is classified into malware by using only the highest weighted output by applying the Attention technique, and learning the RNN output value by Residual CNN again. Experiments on the proposed model showed a Micro-average F1 score of 92% in the validation data set. Experimental results show that the performance of a model capable of learning and classifying arbitrary length data can be verified without special feature extraction and dimension reduction.

• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.327-329
• /
• 2017

개체명 인식(Named Entity Recognition) 시스템은 문서에서 인명(PS), 지명(LC), 단체명(OG)과 같은 개체명을 가지는 단어나 어구를 해당 개체명으로 인식하는 시스템이다. 개체명 인식 시스템을 개발하기 위해 딥러닝 기반의 워드 임베딩(word embedding) 자질과 문장의 형태적 특징 및 기구축 사전(lexicon) 기반의 자질 구성 방법을 제안하고, bi-directional LSTM, CNN, CRF과 같은 모델을 이용하여 구성된 자질을 학습하는 방법을 제안한다. 실험 데이터는 2017 국어 정보시스템 경진대회에서 제공한 2016klpNER 데이터를 이용하였다. 실험은 전체 4258 문장 중 학습 데이터 3406 문장, 검증 데이터 426 문장, 테스트 데이터 426 문장으로 데이터를 나누어 실험을 진행하였다. 실험 결과 본 연구에서 제안하는 모델은 BIO 태깅 방식의 개체 청크 단위 성능 평가 결과 98.9%의 테스트 정확도(test accuracy)와 89.4%의 f1-score를 나타냈다.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.9
• /
• pp.375-384
• /
• 2021

Deep learning techniques have been proven to have high performance in image processing and are applied in various fields. The most widely used methods for validating a deep learning model include a holdout verification method, a k-fold cross verification method, and a bootstrap method. These legacy methods consider the balance of the ratio between classes in the process of dividing the data set, but do not consider the ratio of various features that exist within the same class. If these features are not considered, verification results may be biased toward some features. Therefore, we propose a deep learning model validation method based on data feature coverage for image classification by improving the legacy methods. The proposed technique proposes a data feature coverage that can be measured numerically how much the training data set for training and validation of the deep learning model and the evaluation data set reflects the features of the entire data set. In this method, the data set can be divided by ensuring coverage to include all features of the entire data set, and the evaluation result of the model can be analyzed in units of feature clusters. As a result, by providing feature cluster information for the evaluation result of the trained model, feature information of data that affects the trained model can be provided.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.4
• /
• pp.133-142
• /
• 2021

With the recent advancements of deep learning, companies such as smart home, healthcare, and intelligent transportation systems are utilizing its functionality to provide high-quality services for vehicle detection, emergency situation detection, and controlling energy consumption. To provide reliable services in such sensitive systems, deep learning models are required to have high accuracy. In order to develop a deep learning model for analyzing previously mentioned services, developers should utilize the state of the art deep learning models that have already been verified for higher accuracy. The developers can verify the accuracy of the referenced model by validating the model on the dataset. For this validation, the developer needs structural information to document and apply deep learning models, including metadata such as learning dataset, network architecture, and development environments. In this paper, we propose a description language that represents the network architecture of the deep learning model along with its metadata that are necessary to develop a deep learning model. Through the proposed description language, developers can easily verify the accuracy of the referenced deep learning model. Our experiments demonstrate the application scenario of a deep learning description document that focuses on the license plate recognition for the detection of illegally parked vehicles.