• Proceedings of the Korea Information Processing Society Conference
• /
• 2000.04a
• /
• pp.1062-1066
• /
• 2000

최근 인터넷 사용자의 급속한 증가와 함께 정보 표현 방식도 실세계와 유사한 3차원 공간으로 바뀌고 있다. 본 논문에서는 기존에 개발된 수화 편집기와 수화학습 시스템에서의 수화 동작의 편집 및 수화 학습의 효율을 높이기 위하여 frame 단위의 동작처리를 action 단위로 바꾸어 처리하도록 개선하였다. 개선된 수화 편집기와 수화학습 시스템은 VRML 뷰어와 애플릿간의 자료 이동이 줄어들어 처리 속도의 향상을 가져왔고 자연스러운 수화 동작 구현이 가능하게 되었다.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.135-140
• /
• 2020

사전 학습된 Transformer 기반 언어 모델은 자연어처리 시스템에 적용되었을 시 광범위한 사례에서 큰 폭의 성능 향상을 보여준다. 여기서 사전 학습에 사용되는 언어 모델링 태스크는 비지도 학습에 속하는 기술이기 때문에 상대적으로 데이터의 확보가 쉬운 편이다. 하지만 몇 종의 주류 언어를 제외한 대부분 언어는 활용할 수 있는 언어 자원 자체가 희소하며, 따라서 이러한 사전 학습 기술의 혜택도 누리기 어렵다. 본 연구에서는 이와 같은 상황에서 발생할 수 있는 자원 희소성 문제를 극복하기 위해 이종 언어 간 전이 학습을 이용하는 방법을 제안한다. 본 방법은 언어 자원이 풍부한 언어에서 학습된 Transformer 기반 언어 모델에서 얻은 파라미터 중 재활용 가능한 부분을 이용하여 목표 언어의 모델을 초기화한 후 학습을 진행한다. 또한, 기존 언어와 목표 언어의 차이를 학습하는 역할을 하는 적응층들을 추가하여 이종 언어 간 전이 학습을 돕는다. 제안된 방법을 언어 자원이 희귀한 상황에 대하여 실험해본 결과, 전이 학습을 사용하지 않은 기준 모델 대비 perplexity와 단어 예측의 정확도가 큰 폭으로 향상됨을 확인하였다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.10
• /
• pp.2837-2845
• /
• 2009

This study certified that the mentality training that utilized ICT learning has been working as an important base having much effect on learner's basic attitude on physical education class, improvement of bodily exercise function, and class satisfaction, and that the exercise ability was improved in the scope of speed, form(posture), accuracy(shooting success rate), and adaptability(performance ability). It means it is a much more step-forwarded educational method that the advantages of ICT learning and mentality training at the existing learning method were applied to the reality. Regarding the object of this study, it is a little bit unreasonable to generalize its study results in that it wasn't intended for national unit sampling. Therefore, in the future study, it is necessary to continue to advance the study that its representative-ness was supplemented through the balanced sampling between area and area, and between grade and grade.

• The Journal of Korean Association of Computer Education
• /
• v.5 no.4
• /
• pp.1-8
• /
• 2002

With the development of Internet technology, the construction and spread of network environment increased educational adaption and utilization based on World Wide Web. Learners are educated in the competitive, individual, or collaborative learning structure. Among them, competitive and individual educational methods are criticized for bringing about excessive competition and a lack of cooperation. As a new way of educational method, the interest for the collaborative learning structure was increased. In this perspective, we design and implement a web-based collaborative learning system which is adapted the merit and model of collaboration learning and show that the proposed system improves learning achievement and motivation by experimental study on student groups.

• Journal of IKEEE
• /
• v.23 no.4
• /
• pp.1408-1414
• /
• 2019

Deep artificial neural network with transfer learning is applied to compressed sensing cardiovascular MRI. Transfer learning is a method that utilizes structure, filter kernels, and weights of the network used in prior learning for current learning or application. The transfer learning is useful in accelerating learning speed, and in generalization of the neural network when learning data is limited. From a cardiac MRI experiment, with 8 healthy volunteers, the neural network with transfer learning was able to reduce learning time by a factor of more than five compared to that with standalone learning. Using test data set, reconstructed images with transfer learning showed lower normalized mean square error and better image quality compared to those without transfer learning.

• Journal of KIISE:Software and Applications
• /
• v.34 no.12
• /
• pp.1045-1055
• /
• 2007

Latent topic models are statistical models which automatically captures salient patterns or correlation among features underlying a data collection in a probabilistic way. They are gaining an increased popularity as an effective tool in the application of automatic semantic feature extraction from text corpus, multimedia data analysis including image data, and bioinformatics. Among the important issues for the effectiveness in the application of latent topic models to the massive data set is the efficient learning of the model. The paper proposes an accelerated learning technique for PLSA model, one of the popular latent topic models, by an incremental EM algorithm instead of conventional EM algorithm. The incremental EM algorithm can be characterized by the employment of a series of partial E-steps that are performed on the corresponding subsets of the entire data collection, unlike in the conventional EM algorithm where one batch E-step is done for the whole data set. By the replacement of a single batch E-M step with a series of partial E-steps and M-steps, the inference result for the previous data subset can be directly reflected to the next inference process, which can enhance the learning speed for the entire data set. The algorithm is advantageous also in that it is guaranteed to converge to a local maximum solution and can be easily implemented just with slight modification of the existing algorithm based on the conventional EM. We present the basic application of the incremental EM algorithm to the learning of PLSA and empirically evaluate the acceleration performance with several possible data partitioning methods for the practical application. The experimental results on a real-world news data set show that the proposed approach can accomplish a meaningful enhancement of the convergence rate in the learning of latent topic model. Additionally, we present an interesting result which supports a possible synergistic effect of the combination of incremental EM algorithm with parallel computing.

• Journal of the Korea Society of Computer and Information
• /
• v.7 no.4
• /
• pp.204-210
• /
• 2002

The self-organizing feature map of Kohonen has disadvantage that needs too much input patterns in order to converge into the equilibrium state when it trains. In this paper we proposed the method of improving the convergence speed and rate of self-organizing feature map converting the interaction set into Dynamic Gaussian function. The proposed method Provides us with dynamic Properties that the deviation and width of Gaussian function used as an interaction function are narrowed in proportion to learning times and learning rates that varies according to topological position from the winner neuron. In this Paper. we proposed the method of improving the convergence rate and the degree of self-organizing feature map.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.5
• /
• pp.677-682
• /
• 2010

This paper addresses a method of estimating roughly the human pose by comparing Haar-wavelet value which is learned in face detection technology using AdaBoost algorithm. We also presents its application to face recognition. The learned weak classifier is used to a Haar-wavelet robust to each pose's feature by comparing the coefficients during the process of face detection. The Mahalanobis distance is used to measure the matching degree in Haar-wavelet selection. When a facial image is detected using the selected Haar-wavelet, the pose is estimated. The proposed pose estimation can be used to improve face recognition speed. Experiments are conducted to evaluate the performance of the proposed method for pose estimation.

• Proceedings of the IEEK Conference
• /
• 2000.09a
• /
• pp.435-438
• /
• 2000

본 논문에서는 웨이블렛에 기반한 FLD(Fisher Linear Discriminant) 방법을 제안한다. 본 논문은 얼굴인식에 대한 속도와 정확성을 다룬다. 128×128의 해상도를 가진 영상은 웨이블렛 변환을 통해 16×16의 부영상들로 분해된 후에, 저대역과 중대역에 해당하는 두 개의 부영상을 사용하여 학습과 인식을 한다. 실험 결과, 제안된 방법은 기존의 FLD 방법의 인식률을 유지하며, 보다 더 빠른 속도를 가진다. 우리의 실험에서는 약 6배의 속도 향상을 보인다.

• Journal of Practical Engineering Education
• /
• v.14 no.1
• /
• pp.219-224
• /
• 2022

Q-Learning is a technique widely used as a basic algorithm for reinforcement learning. Q-Learning trains the agent in the direction of maximizing the reward through the greedy action that selects the largest value among the rewards of the actions that can be taken in the current state. In this paper, we studied a policy that can speed up agent training using Q-Learning in Frozen Lake 8×8 grid environment. In addition, the training results of the existing algorithm of Q-learning and the algorithm that gave the attribute 'direction' to agent movement were compared. As a result, it was analyzed that the Q-Learning policy proposed in this paper can significantly increase both the accuracy and training speed compared to the general algorithm.