• Journal of Satellite, Information and Communications
• /
• v.6 no.2
• /
• pp.78-83
• /
• 2011

In this paper, we propose a computationally efficient early stopping method to reduce the average number of iterations. The conventional early stopping methods have too much computational complexity to compute the stopping criterion. Thus, only the hard decision based early stopping method is suitable to realize the hardware of LDPC decoder. However, this method also can increase the computational complexity of LDPC decoder. The proposed method can effectively reduce the computational complexity of stopping criterion as we do not compute hard decision, and we combine the stopping criterion with horizontal shuffling scheduling decoding scheme. The simulation results show that a new early stopping method achieves acceptable bit error rate performance also reduces the average number of iterations.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.11
• /
• pp.2095-2101
• /
• 2008

In this paper, we consider a performance improvement of neural network for predicting defect size of steam generator tube using early stopping. Usually, neural network is trained until MSE becomes less than a prescribed error goal. The smaller the error goal, the greater the prediction performance for the trained data. However, as the error goal is decreased, an over fitting is likely to start during supervised training of a neural network, which usually deteriorates the generalization performance. We propose that, for the prediction of an axisymmetric defect size, early stopping can be used to avoid the over-fitting. Through various experiments on the axisymmetric defect samples, we found that the difference bet ween the prediction error of neural network based on early stopping and that of ideal neural network is reasonably small. This indicates that the error goal used for neural network training for the prediction of defect size can be efficiently selected by early stopping.

• Journal of information and communication convergence engineering
• /
• v.21 no.3
• /
• pp.198-207
• /
• 2023

Responding to changes in artificial intelligence models and the data environment is crucial for increasing data-learning accuracy and inference stability of industrial applications. A learning model that is overfitted to specific training data leads to poor learning performance and a deterioration in flexibility. Therefore, an early stopping technique is used to stop learning at an appropriate time. However, this technique does not consider the homogeneity and independence of the data collected by heterogeneous nodes in a differential network environment, thus resulting in low learning accuracy and degradation of system performance. In this study, the generalization performance of neural networks is maximized, whereas the effect of the homogeneity of datasets is minimized by achieving an accuracy of 99.7%. This corresponds to a decrease in delay time by a factor of 2.33 and improvement in performance by a factor of 2.5 compared with the conventional method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.6A
• /
• pp.537-543
• /
• 2007

In this paper, we propose a new stopping criterion for the turbo processing (Turbo-BLAST) of MIMO-OFDM system. To reduce the high computational complexity of turbo-BLAST, it is desirable to lessen the outer-loop iteration number. In a system such as IEEE 802.16e Mobile WiMax, no CRC bits are available except the last encoding packet of a transmitted burst, so early stopping criteria without the help of CRC bits are needed. The proposed criterion counts the sign differences between received parity bits and the re-encoded parity bits from received information bits. With the tail-biting code which is accepted for IEEE 802.16e, a method that the re-encoder operates at half complexity is also proposed. Computer simulations show that the proposed stopping criterion approaches the performance of GENIE aided criterion with less average number of iterations than the other early stopping criteria.

• Journal of Advanced Information Technology and Convergence
• /
• v.8 no.2
• /
• pp.119-127
• /
• 2018

Training a very large deep neural network can be painfully slow and prone to overfitting. Many researches have done for overcoming the problem. In this paper, a combination of early stopping and ADAM based deep neural network was presented. This form of deep network is useful for handling the big data because it automatically stop the training before overfitting occurs. Also generalization ability is better than pure deep neural network model.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.45 no.5
• /
• pp.117-125
• /
• 2003

Since the recorded information used for operation of a catchment modelling system contain errors that influence the calibration of catchment modelling system control parameter values, the accurate estimation of these parameters is difficult. Despite these influences, existing traditional calibration approaches focus only on achieving the best "curve fitting" between simulated and recorded data, and not on generic evaluation of control parameter values. This paper introduces an Early Stopping Technique which is aimed at avoiding the procedure of curve-fitting through monitoring improvements in the objective function used for assessing the optimal parameter set. Application of this approach to the calibration of SWMM (Storm Water Management Model) on the Centennial Park catchment in Sydney, Australia is outlined. outlined.

• The KIPS Transactions:PartD
• /
• v.8D no.2
• /
• pp.175-180
• /
• 2001

This paper presents the optimal fraction of validation set to obtain a prediction accuracy of software failure count or failure time in the future by a neuro-fuzzy system. Given a fixed amount of training data, the most popular effective approach to avoiding underfitting and overfitting is early stopping, and hence getting optimal generalization. But there is unresolved practical issues : How many data do you assign to the training and validation set\ulcorner Rules of thumb abound, the solution is acquired by trial-and-error and we spend long time in this method. For the sake of optimal fraction of validation set, the variant specific fraction for the validation set be provided. It shows that minimal fraction of the validation data set is sufficient to achieve good next-step prediction. This result can be considered as a practical guideline in a prediction of software reliability by neuro-fuzzy system.

• Bulletin of the Korean Mathematical Society
• /
• v.48 no.4
• /
• pp.791-812
• /
• 2011

This is a survey on American options. An American option allows its owner the privilege of early exercise, whereas a European option can be exercised only at expiration. Because of this early exercise privilege American option pricing involves an optimal stopping problem; the price of an American option is given as a free boundary value problem associated with a Black-Scholes type partial differential equation. Up until now there is no simple closed-form solution to the problem, but there have been a variety of approaches which contribute to the understanding of the properties of the price and the early exercise boundary. These approaches typically provide numerical or approximate analytic methods to find the price and the boundary. Topics included in this survey are early approaches(trees, finite difference schemes, and quasi-analytic methods), an analytic method of lines and randomization, a homotopy method, analytic approximation of early exercise boundaries, Monte Carlo methods, and relatively recent topics such as model uncertainty, backward stochastic differential equations, and real options. We also provide open problems whose answers are expected to contribute to American option pricing.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.6
• /
• pp.51-58
• /
• 2020

The blind people have difficulty living an independent clothing life. The furniture and home appliance are adding AI or IoT with the recent growth of the smart appliance market. To support the independent clothing life of the blind, this paper suggests a smart wardrobe with closet control function, voice recognition function and clothes information recognition using CNN algorithm. The number of layers of the model was changed and Maxpooling was adjusted to create the model to increase accuracy in the process of recognizing clothes. Early Stopping Callback option is applied to ensure learning accuracy when creating a model. We added Dropout to prevent overfitting. The final model created by this process can be found to have 80 percent accuracy in clothing recognition.

• The Korean Journal of Pain
• /
• v.24 no.2
• /
• pp.112-114
• /
• 2011

Due to its complex pathophysiology and wide spectrum of clinical manifestations, the diagnosis of CRPS is often missed in the early stage by primary care physicians. After being treated by a primary care physician for 5 months for chronic cellulitis, a 16-year-old girl was referred to our hospital with features of type-1 CRPS of the right upper extremity. Inability to diagnose early caused prolonged suffering to the girl with all the consequence of CRPS. The patient responded well with marked functional recovery from multimodal therapy. Ability to distinguish CRPS from other pain conditions, referral for specialty care at the appropriate time and full awareness of this condition and its clinical features among various healthcare professionals are essential in reducing patient suffering and stopping its progression towards difficult-to-treat situations.