• Journal of the Korea Institute of Military Science and Technology
• /
• v.22 no.1
• /
• pp.49-59
• /
• 2019

Generative adversarial networks(GANs) have received great attention in the machine learning field for their capacity to model high-dimensional and complex data distribution implicitly and generate new data samples from the model distribution. This paper investigates the model training methodology, architecture, and various applications of generative adversarial networks. Experimental evaluation is also conducted for generating synthetic image dataset for defense using two types of GANs. The first one is for military image generation utilizing the deep convolutional generative adversarial networks(DCGAN). The other is for visible-to-infrared image translation utilizing the cycle-consistent generative adversarial networks(CycleGAN). Each model can yield a great diversity of high-fidelity synthetic images compared to training ones. This result opens up the possibility of using inexpensive synthetic images for training neural networks while avoiding the enormous expense of collecting large amounts of hand-annotated real dataset.

• Progress in Medical Physics
• /
• v.31 no.3
• /
• pp.111-123
• /
• 2020

Deep learning (DL) is a subset of machine learning and artificial intelligence that has a deep neural network with a structure similar to the human neural system and has been trained using big data. DL narrows the gap between data acquisition and meaningful interpretation without explicit programming. It has so far outperformed most classification and regression methods and can automatically learn data representations for specific tasks. The application areas of DL in radiation oncology include classification, semantic segmentation, object detection, image translation and generation, and image captioning. This article tries to understand what is the potential role of DL and what can be more achieved by utilizing it in radiation oncology. With the advances in DL, various studies contributing to the development of radiation oncology were investigated comprehensively. In this article, the radiation treatment process was divided into six consecutive stages as follows: patient assessment, simulation, target and organs-at-risk segmentation, treatment planning, quality assurance, and beam delivery in terms of workflow. Studies using DL were classified and organized according to each radiation treatment process. State-of-the-art studies were identified, and the clinical utilities of those researches were examined. The DL model could provide faster and more accurate solutions to problems faced by oncologists. While the effect of a data-driven approach on improving the quality of care for cancer patients is evidently clear, implementing these methods will require cultural changes at both the professional and institutional levels. We believe this paper will serve as a guide for both clinicians and medical physicists on issues that need to be addressed in time.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.4
• /
• pp.169-178
• /
• 2022

Morphemes are most primitive units in a language that lose their original meaning when segmented into smaller parts. In Korean, a sentence is a sequence of eojeols (words) separated by spaces. Each eojeol comprises one or more morphemes. Korean morphological analysis (KMA) is to divide eojeols in a given Korean sentence into morpheme units. It also includes assigning appropriate part-of-speech(POS) tags to the resulting morphemes. KMA is one of the most important tasks in Korean natural language processing (NLP). Improving the performance of KMA is closely related to increasing performance of Korean NLP tasks. Recent research on KMA has begun to adopt the approach of machine translation (MT) models. MT is to convert a sequence (sentence) of units of one domain into a sequence (sentence) of units of another domain. Neural machine translation (NMT) stands for the approaches of MT that exploit neural network models. From a perspective of MT, KMA is to transform an input sequence of units belonging to the eojeol domain into a sequence of units in the morpheme domain. In this paper, we propose a deep learning model for KMA. The backbone of our model is based on the BERT-fused model which was shown to achieve high performance on NMT. The BERT-fused model utilizes Transformer, a representative model employed by NMT, and BERT which is a language representation model that has enabled a significant advance in NLP. The experimental results show that our model achieves 98.24 F1-Score.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.535-540
• /
• 2022

Out of Vocabulary(OOV) 문제는 인공신경망 기계번역(Neural Machine Translation, NMT)에서 빈번히 제기되어 왔다. 이를 해결하기 위해, 기존에는 단어를 효율적인 압축할 수 있는 Byte Pair Encoding(BPE)[1]이 대표적으로 이용되었다. 하지만 BPE는 빈도수를 기반으로 토큰화가 진행되는 결정론적 특성을 취하고 있기에, 다양한 문장에 관한 일반화된 분절 능력을 함양하기 어렵다. 이를 극복하기 위해 최근 서브 워드를 정규화하는 방법(Subword Regularization)이 제안되었다. 서브 워드 정규화는 동일한 단어 안에서 발생할 수 있는 다양한 분절 경우의 수를 고려하도록 설계되어 다수의 실험에서 우수한 성능을 보였다. 그러나 분류 작업, 특히 한국어를 대상으로 한 분류에 있어서 서브 워드 정규화를 적용한 사례는 아직까지 확인된 바가 없다. 이를 위해 본 논문에서는 서브 워드 정규화를 대표하는 두 가지 방법인 유니그램 기반 서브 워드 정규화[2]와 BPE-Dropout[3]을 이용해 한국어 분류 문제에 대한 서브 워드 정규화의 효과성을 제안한다. NMT 뿐만 아니라 분류 문제 역시 단어의 구성성 및 그 의미를 파악하는 것은 각 문장이 속하는 클래스를 결정하는데 유의미한 기여를 한다. 더불어 서브 워드 정규화는 한국어의 문장 구성 요소에 관해 폭넓은 인지능력을 함양할 수 있다. 해당 방법은 본고에서 진행한 한국어 분류 과제 실험에서 기존 BPE 대비 최대 4.7% 높은 성능을 거두었다.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.441-445
• /
• 2020

토큰화(Tokenization)는 사람이 작성한 자연어 문장을 기계가 잘 이해할 수 있도록 최소 단위인 토큰으로 분리하는 작업을 말하여, 이러한 토큰화는 자연어처리 전반적인 태스크들의 전처리에 필수적으로 사용되고 있다. 최근 자연어처리 분야에서 높은 성능을 보이며, 다양한 딥러닝 모델에 많이 활용되고 있는 SentencePiece 토큰화는 여러 단어에서 공통적으로 출현하는 부분단어들을 기준으로, BPE 알고리즘을 이용하여 문장을 압축 표현하는 토큰화 방법이다. 본 논문에서는 한국어 기반 특허 문헌의 초록 자연어 데이터를 기반으로 SentencePiece를 비롯한 여러 토큰화 방법에 대하여 소개하며, 해당 방법을 응용한 기계번역 (Neural Machine Translation) 태스크를 수행하고, 토큰화 방법별 비교 평가를 통해 특허 분야 자연어 데이터에 최적화된 토큰화 방법을 제안한다. 그리고 본 논문에서 제안한 방법을 사용하여 특허 초록 한-영 기계번역 태스크에서 성능이 향상됨을 보였다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.4C
• /
• pp.494-504
• /
• 2004

The researches about gaze detection have been much developed with many applications. Most previous researches only rely on image processing algorithm, so they take much processing time and have many constraints. In our work, we implement it with a computer vision system setting a IR-LED based single camera. To detect the gaze position, we locate facial features, which is effectively performed with IR-LED based camera and SVM(Support Vector Machine). When a user gazes at a position of monitor, we can compute the 3D positions of those features based on 3D rotation and translation estimation and affine transform. Finally, the gaze position by the facial movements is computed from the normal vector of the plane determined by those computed 3D positions of features. In addition, we use a trained neural network to detect the gaze position by eye's movement. As experimental results, we can obtain the facial and eye gaze position on a monitor and the gaze position accuracy between the computed positions and the real ones is about 4.2 cm of RMS error.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.2
• /
• pp.79-88
• /
• 2004

Gaze detection is to locate the position on a monitor screen where a user is looking by computer vision. Gaze detection systems have numerous fields of application. They are applicable to the man-machine interface for helping the handicapped to use computers and the view control in three dimensional simulation programs. In our work, we implement it with a computer vision system setting a IR-LED based single camera. To detect the gaze position, we locate facial features, which is effectively performed with IR-LED based camera and SVM(Support Vector Machine). When a user gazes at a position of monitor, we can compute the 3D positions of those features based on 3D rotation and translation estimation and affine transform. Finally, the gaze position by the facial movements is computed from the normal vector of the plane determined by those computed 3D positions of features. In addition, we use a trained neural network to detect the gaze position by eye's movement. As experimental results, we can obtain the facial and eye gaze position on a monitor and the gaze position accuracy between the computed positions and the real ones is about 4.8 cm of RMS error.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.1
• /
• pp.27-36
• /
• 2019

In this paper, we propose a method to find the most similar answer to the user's response from the question-answer database in order to avoid generating a redundant question in retrieval-based automatic question generation system. As a question of the most similar answer to user's response may already be known to the user, the question should be removed from a set of question candidates. A similarity detector calculates a similarity between two answers by utilizing the same words, paraphrases, and sentential meanings. Paraphrases can be acquired by building a phrase table used in a statistical machine translation. A sentential meaning's similarity of two answers is calculated by an attention-based convolutional neural network. We evaluate the accuracy of the similarity detector on an evaluation set with 100 answers, and can get the 71% Mean Reciprocal Rank (MRR) score.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.4
• /
• pp.163-170
• /
• 2019

Korea language has postpositions such as eul, reul, yi, ga, wa, and gwa, which are attached to nouns and add meaning to the sentence. When foreign notations or abbreviations are included in sentences, the appropriate postposition for the pronunciation of the foreign words may not be used. Sometimes, for natural expression of the sentence, two postpositions are used with one in parentheses as in "eul(reul)" so that both postpositions can be acceptable. This study finds examples of using unnatural postpositions when foreign words are included in Korean sentences and proposes a method for using natural postpositions by learning the final consonant pronunciation of nouns. The proposed method uses a recurrent neural network model to naturally express postpositions connected to foreign words. Furthermore, the proposed method is proven by learning and testing with the proposed method. It will be useful for composing perfect sentences for machine translation by using natural postpositions for English abbreviations or new foreign words included in Korean sentences in the future.

• Journal of the Korea Convergence Society
• /
• v.12 no.11
• /
• pp.35-43
• /
• 2021

Recently, there has been a growing interest in zero-shot cross-lingual transfer, which leverages cross-lingual language models (CLLMs) to perform downstream tasks that are not trained in a specific language. In this paper, we point out the limitations of the data-centric aspect of quality estimation (QE), and perform zero-shot cross-lingual transfer even in environments where it is difficult to construct QE data. Few studies have dealt with zero-shots in QE, and after fine-tuning the English-German QE dataset, we perform zero-shot transfer leveraging CLLMs. We conduct comparative analysis between various CLLMs. We also perform zero-shot transfer on language pairs with different sized resources and analyze results based on the linguistic characteristics of each language. Experimental results showed the highest performance in multilingual BART and multillingual BERT, and we induced QE to be performed even when QE learning for a specific language pair was not performed at all.