• Journal of Institute of Control, Robotics and Systems
• /
• v.5 no.2
• /
• pp.174-179
• /
• 1999

In this paper, we propose an optimal algorithm for finding the shortest connection of two kinds of parts to pairs. If total part numbers are of size N, then there are order 2ㆍ(N/2)$^{N}$ possible solutions, of which we want the one that minimizes the energy function. The appropriate dynamic rule and parameters used in network are proposed by a new energy function which is minimized when 3-constraints are satisfied. This dynamic nile has three important parameters, an enhancement variable connected to pairs, a normalized distance term and a time variable. The enhancement variable connected to pairs have to a perfect connection of two kinds of parts to pairs. The normalized distance term get rids of a unstable states caused by the change of total part numbers. And the time variable removes the un-optimal connection in the case of distance constraint and the wrong or not connection of two kinds of parts to pairs. First of all, we review the theoretical basis for Hopfield model and present a new energy function. Then, the connection matrix and the offset bias created by a new energy function and used in dynamic nile are shown. Finally, we show examples through computer simulation with 20, 30 and 40 parts and discuss the stability and feasibility of the resultant solutions for the proposed connection algorithm.m.

• Phonetics and Speech Sciences
• /
• v.13 no.1
• /
• pp.45-51
• /
• 2021

End-to-end (E2E) automatic speech recognition (ASR) has achieved promising performance gains with the introduced self-attention network, Transformer. However, due to training time and the number of hyperparameters, finding the optimal hyperparameter set is computationally expensive. This paper investigates the impact of hyperparameters in the Transformer network to answer two questions: which hyperparameter plays a critical role in the task performance and training speed. The Transformer network for training has two encoder and decoder networks combined with Connectionist Temporal Classification (CTC). We have trained the model with Wall Street Journal (WSJ) SI-284 and tested on devl93 and eval92. Seventeen hyperparameters were selected from the ESPnet training configuration, and varying ranges of values were used for experiments. The result shows that "num blocks" and "linear units" hyperparameters in the encoder and decoder networks reduce Word Error Rate (WER) significantly. However, performance gain is more prominent when they are altered in the encoder network. Training duration also linearly increased as "num blocks" and "linear units" hyperparameters' values grow. Based on the experimental results, we collected the optimal values from each hyperparameter and reduced the WER up to 2.9/1.9 from dev93 and eval93 respectively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.1
• /
• pp.334-349
• /
• 2022

The effectiveness of Wi-Fi networks is greatly influenced by the optimization of contention window (CW) parameters. Unfortunately, the conventional approach employed by IEEE 802.11 wireless networks is not scalable enough to sustain consistent performance for the increasing number of stations. Yet, it is still the default when accessing channels for single-users of 802.11 transmissions. Recently, there has been a spike in attempts to enhance network performance using a machine learning (ML) technique known as reinforcement learning (RL). Its advantage is interacting with the surrounding environment and making decisions based on its own experience. Deep RL (DRL) uses deep neural networks (DNN) to deal with more complex environments (such as continuous state spaces or actions spaces) and to get optimum rewards. As a result, we present a new approach of CW control mechanism, which is termed as contention window threshold (CW_Threshold). It uses the DRL principle to define the threshold value and learn optimal settings under various network scenarios. We demonstrate our proposed method, known as a smart exponential-threshold-linear backoff algorithm with a deep Q-learning network (SETL-DQN). The simulation results show that our proposed SETL-DQN algorithm can effectively improve the throughput and reduce the collision rates.

• Journal of Digital Contents Society
• /
• v.19 no.6
• /
• pp.1161-1167
• /
• 2018

The development of machine learning has enabled machines to perform delicate tasks that only humans could do, and thus many companies have introduced machine learning based translators. Existing translators have good performances but they have problems in number translation. The translators often mistranslate numbers when the input sentence includes a large number. Furthermore, the output sentence structure completely changes even if only one number in the input sentence changes. In this paper, first, we optimized a neural machine translation model architecture that uses bidirectional RNN, LSTM, and the attention mechanism through data cleansing and changing the dictionary size. Then, we implemented a number-processing algorithm specialized in number translation and applied it to the neural machine translation model to solve the problems above. The paper includes the data cleansing method, an optimal dictionary size and the number-processing algorithm, as well as experiment results for translation performance based on the BLEU score.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.5 no.4
• /
• pp.41-55
• /
• 1995

In this paper, an optimal idenfication method using fuzzy-neural networks is proposed for modeling of nonlinear complex systems. The proposed fuzzy-neural modeling implements system structure and parameter identification using the intelligent schemes together wlth optimization theory, linguistic fuzzy implication rules, and neural networks(NNs) from input and output data of processes. Inference type for this fuzzy-neural modeling is presented as simplified inference. To obtain optimal model, the learning rates and momentum coefficients of fuzzy-neural networks(FNNs) are tuned automatically using improved modified complex method and modified learning algorithm. For the purpose of its application to nonlinear processes, data for route choice of traffic problems and those for activateti sluge process of sewage treatment system are used for the purpose of evaluating the performance of the proposed fuzzy-neural network modeling. The results show that the proposed method can produce the intelligence model with higher accuracy than other works achieved previously.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.06a
• /
• pp.612-618
• /
• 1998

The success of a fuzzy logic control system solving any given problem critically depends on the architecture of th network. Various attempts have been made in optimizing its structure its structure using genetic algorithm automated designs. In a regular genetic algorithm , a difficulty exists which lies in the encoding of the problem by highly fit gene combinations of a fixed-length. This paper presents a new approach to structurally optimized designs of a fuzzy model. We use a messy genetic algorithm, whose main characteristics is the variable length of chromosomes. A messy genetic algorithms used to obtain structurally optimized fuzzy models. Structural optimization is regarded important before neural network based learning is switched into. We have applied the method to the exampled of a cart-pole balancing.

• Earthquakes and Structures
• /
• v.24 no.2
• /
• pp.111-126
• /
• 2023

Highly reliable and versatile methods artificial intelligence (AI) have found multiple application in the different fields of science, engineering and health care system. In the present study, we aim to utilize AI method to investigated vibrations in the human leg bone. In this regard, the bone geometry is simplified as a thick cylindrical shell structure. The deep neural network (DNN) is selected for prediction of natural frequency and critical buckling load of the bone cylindrical model. Training of the network is conducted with results of the numerical solution of the governing equations of the bone structure. A suitable optimization algorithm is selected for minimizing the loss function of the DNN. Generalized differential quadrature method (GDQM), and Hamilton's principle are used for solving and obtaining the governing equations of the system. As well as this, in the results section, with the aid of AI some predictions for improving the behaviors of the various sport systems will be given in detail.

• International Journal of Computer Science & Network Security
• /
• v.22 no.7
• /
• pp.91-102
• /
• 2022

With the evolution of wireless sensor network (WSN) technology, the routing policy has foremost importance in the Internet of Things (IoT). A systematic routing policy is one of the primary mechanics to make certain the precise and robust transmission of wireless sensor networks in an energy-efficient manner. In an IoT environment, WSN is utilized for controlling services concerning data like, data gathering, sensing and transmission. With the advantages of IoT potentialities, the traditional routing in a WSN are augmented with decision-making in an energy efficient manner to concur finer optimization. In this paper, we study how to combine IoT-based deep learning classifier with routing called, Kriging Regressive Deep Belief Neural Learning (KR-DBNL) to propose an efficient data packet routing to cope with scalability issues and therefore ensure robust data packet transmission. The KR-DBNL method includes four layers, namely input layer, two hidden layers and one output layer for performing data transmission between source and destination sensor node. Initially, the KR-DBNL method acquires the patient data from different location. Followed by which, the input layer transmits sensor nodes to first hidden layer where analysis of energy consumption, bandwidth consumption and light intensity are made using kriging regression function to perform classification. According to classified results, sensor nodes are classified into higher performance and lower performance sensor nodes. The higher performance sensor nodes are then transmitted to second hidden layer. Here high performance sensor nodes neighbouring sensor with higher signal strength and frequency are selected and sent to the output layer where the actual data packet transmission is performed. Experimental evaluation is carried out on factors such as energy consumption, packet delivery ratio, packet loss rate and end-to-end delay with respect to number of patient data packets and sensor nodes.

• Journal of Korea Multimedia Society
• /
• v.6 no.1
• /
• pp.136-147
• /
• 2003

To achieve an optimal performance for a given problem, we need an integrative process of the parameter optimization via learning and the structure optimization via model selection. In this paper, we propose an efficient optimization method for improving generalization performance by considering the property of each sub-method and by combining them with common theoretical properties. First, weight parameters are optimized by natural gradient teaming with adaptive regularization, which uses a diverse error function. Second, the network structure is optimized by eliminating unnecessary parameters with natural pruning. Through iterating these processes, candidate models are constructed and evaluated based on the Bayesian Information Criterion so that an optimal one is finally selected. Through computational experiments on benchmark problems, we confirm the weight parameter and structure optimization performance of the proposed method.

• The KSFM Journal of Fluid Machinery
• /
• v.13 no.5
• /
• pp.43-53
• /
• 2010

This study investigates a design optimization of a rotating two-pass rectangular cooling channel with staggered arrays of pin-fins. The radial basis neural network method is used as an optimization technique with Reynolds-averaged Navier-Stokes analysis of fluid flow and heat transfer with shear stress transport turbulent model. The ratio of the diameter to height of the pin-fins and the ratio of the streamwise spacing between the pin-fins to height of the pin-fin are selected as design variables. The optimization problem has been defined as a minimization of the objective function, which is defined as a linear combination of heat transfer related term and friction loss related term with a weighting factor. Results are presented for streamlines, velocity vector fields, and contours of Nusselt numbers, friction coefficients, and turbulent kinetic energy. These results show how fluid flow in a two-pass square cooling channel evolves a converted secondary flows due to Coriolis force, staggered arrays of pin-fins, and a $180^{\circ}$ turn region. These results describe how the fluid flow affects surface heat transfer. The Coriolis force induces heat transfer discrepancy between leading and trailing surfaces, having higher Nusselt number on the leading surface in the second pass while having lower Nusselt number on the trailing surface. Dean vortices generated in $180^{\circ}$ turn region augment heat transfer in the turning region and in the upstream region of the second pass. As the result of optimization, in comparison with the reference geometry, thermal performance of the optimum geometry shows the improvement by 30.5%. Through the optimization, the diameter of pin-fin increased by 14.9% and the streamwise distance between pin-fins increased by 32.1%. And, the value of objective function decreased by 18.1%.