• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.115-118
• /
• 2020

딥러닝은 자연어처리 분야에서 우수한 성능을 보이고 있다. 하지만 우수한 성능을 달성하려면 많은 학습 데이터와 오랜 학습 시간이 필요하다. 우리는 딥러닝과 기호 규칙을 함께 사용하는 뉴럴-심볼릭 방법을 이용하여 딥러닝만으로 학습한 모델의 성능을 능가하는 방법을 제안한다. 딥러닝의 한계를 극복하기 위해서 관계추출에서 규칙 결과와 딥러닝 결과와의 불일치도를 추가한 구조를 설계하였다. 제안한 구조는 한국어 데이터에 대해서 우수한 성능을 보였으며, 빠른 성능 수렴이 이루어지는 것을 확인하였다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.7
• /
• pp.862-868
• /
• 2007

The planar motion of the index finger in general human hands is usually implemented by the actuation of three joints. This task requires a technique to determine the joint combination for each fingertip position which is well-known as the inverse kinematics problem in robotics. Especially, it is an essential work for grasping and manipulation tasks by robotic and humanoid fingers. In this paper, an intelligent neural learning scheme for solving such inverse kinematics is presented. Specifically, a multi-layered neural network is utilized for effective inverse kinematics, where a dynamic neural learning algorithm is employed for fast learning. Also, a bio-mimetic feature of general human fingers is incorporated to the learning scheme. The usefulness of the proposed approach is verified by simulations.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.64-67
• /
• 2020

최근 다양한 분야에서 딥 러닝 기반의 많은 연구가 진행되고 있으며 이에 따라 딥 러닝 모델의 경량화를 통해 제한된 메모리를 가진 하드웨어에 올릴 수 있는 경량화 된 딥 뉴럴 네트워크(DNN)를 개발하는 연구도 활발해졌다. 이에 본 논문은 주파수 영역에서의 군집화 기반 계층별 딥 뉴럴 네트워크 압축을 제안한다. 이산 코사인 변환, 양자화, 군집화, 적응적 엔트로피 코딩 과정을 각 모델의 계층에 순차적으로 적용하여 DNN이 차지하는 메모리를 줄인다. 제안한 알고리즘을 통해 VGG16을 손실률은 1% 미만의 손실에서 전체 가중치를 3.98%까지 압축, 약 25배가량 경량화 할 수 있었다.

• Journal of Convergence for Information Technology
• /
• v.11 no.5
• /
• pp.30-37
• /
• 2021

Various devices are connected to the Internet, and attacks using the Internet are occurring. Among such attacks, there are attacks that use malicious URLs to make users access to wrong phishing sites or distribute malicious viruses. Therefore, how to detect such malicious URL attacks is one of the important security issues. Among recent deep learning technologies, neural networks are showing good performance in image recognition, speech recognition, and pattern recognition. This neural network can be applied to research that analyzes and detects patterns of malicious URL characteristics. In this paper, performance analysis according to various parameters was performed on a method of detecting malicious URLs using neural networks. In this paper, malicious URL detection performance was analyzed while changing the activation function, learning rate, and neural network structure. The experimental data was crawled by Alexa top 1 million and Whois to build the data, and the machine learning library used TensorFlow. As a result of the experiment, when the number of layers is 4, the learning rate is 0.005, and the number of nodes in each layer is 100, the accuracy of 97.8% and the f1 score of 92.94% are obtained.

• Journal of Digital Convergence
• /
• v.15 no.3
• /
• pp.201-209
• /
• 2017

This paper is a smartphone object recognition system using hierarchical convolutional neural network. The overall configuration is a method of communicating object information to the smartphone by matching the collected data by connecting the smartphone and the server and recognizing the object to the convergence neural network in the server. It is also compared to a hierarchical convolutional neural network and a fractional convolutional neural network. Hierarchical convolutional neural networks have 88% accuracy, fractional convolutional neural networks have 73% accuracy and 15%p performance improvement. Based on this, it shows possibility of expansion of T-Commerce market connected with smartphone and broadcasting media.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.529-534
• /
• 2020

인공신경망 기반 기계번역(Neural Machine Translation, NMT)이란 딥러닝(Deep learning)을 이용하여 출발 언어의 문장을 도착 언어 문장으로 번역해주는 시스템을 일컫는다. NMT는 종단간 학습(end-to-end learning)을 이용하여 기존 기계번역 방법론의 성능을 앞지르며 기계번역의 주요 방법론으로 자리잡게 됐다. 이러한 발전에도 불구하고 여전히 개체(entity), 또는 전문 용어(terminological expressions)의 번역은 미해결 과제로 남아있다. 개체나 전문 용어는 대부분 명사로 구성되는데 문장 내 명사는 주체, 객체 등의 역할을 하는 중요한 요소이므로 이들의 정확한 번역이 문장 전체의 번역 성능 향상으로 이어질 수 있다. 따라서 본 논문에서는 지식그래프(Knowledge Graph)를 이용하여 심볼릭 지식을 NMT와 결합한 뉴럴심볼릭 방법론을 제안한다. 또한 지식그래프를 활용하여 NMT의 성능을 높인 선행 연구 방법론을 한영 기계번역에 이용할 수 있도록 구조를 설계한다.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.254-256
• /
• 2020

자연언어처리 문제에서 딥러닝 모델이 좋은 성능을 보이고 있고 딥러닝 결과는 구조화된 결과를 내놓는 경우가 많다. 딥러닝 모델 결과가 구조적인 형태를 가지는 경우 후처리 통해 특정 구조에 맞는 제약을 가해주는 경우가 일반적이다. 본 논문에서는 이러한 제약을 규칙에 기반하지 않고 직접 학습을 통해 얻고자 하였다.

• Smart Media Journal
• /
• v.7 no.3
• /
• pp.16-21
• /
• 2018

This paper presents the design and implementation of a deep learning-based automated weed detector on onion fields. The system is based on a Convolutional Neural Network that specifically selects proposed regions. The detector initiates training with a dataset taken from agricultural onion fields, after which candidate regions with very high probability of suspicion are considered weeds. Non-maximum suppression helps preserving the less overlapped bounding boxes. The dataset collected from different onion farms is evaluated with the proposed classifier. Classification accuracy is about 99% for the dataset, indicating the proposed method's superior performance with regard to weed detection on the onion fields.

• Korean Journal of Cognitive Science
• /
• v.28 no.4
• /
• pp.299-314
• /
• 2017

Predicting compound-protein interactions in-silico is significant for the drug discovery. In this paper, we propose an scalable machine learning model to predict compound-protein interaction. The key idea of this scalable machine learning model is the architecture of pairwise neural network model and feature embedding method from the raw data, especially for protein. This method automatically extracts the features without additional knowledge of compound and protein. Also, the pairwise architecture elevate the expressiveness and compact dimension of feature by preventing biased learning from occurring due to the dimension and type of features. Through the 5-fold cross validation results on large scale database show that pairwise neural network improves the performance of predicting compound-protein interaction compared to previous prediction models.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.137-142
• /
• 2024

Recently, deep learning technology has become those methods as de facto standards in the area of medical data representation. But, deep learning inherently requires a large amount of training data, which poses a challenge for its direct application in the medical field where acquiring large-scale data is not straightforward. Additionally, brain signal modalities also suffer from these problems owing to the high variability. Research has focused on designing deep neural network structures capable of effectively extracting spectro-spatio-temporal characteristics of brain signals, or employing self-supervised learning methods to pre-learn the neurophysiological features of brain signals. This paper analyzes methodologies used to handle small-scale data in emerging fields such as brain-computer interfaces and brain signal-based state prediction, presenting future directions for these technologies. At first, this paper examines deep neural network structures for representing brain signals, then analyzes self-supervised learning methodologies aimed at efficiently learning the characteristics of brain signals. Finally, the paper discusses key insights and future directions for deep learning-based brain signal analysis.