• Journal of IKEEE
• /
• v.23 no.4
• /
• pp.1314-1320
• /
• 2019

In this paper, we propose a new method for optimizing bit operations and applying them to DNN(Deep Neural Network) in software environment. As a method for this, we propose a packing function for bitwise optimization and a masking matrix multiplication operation for application to DNN. The packing function converts 32-bit real value to 2-bit quantization value through threshold comparison operation. When this sequence is over, four 32-bit real values are changed to one 8-bit value. The masking matrix multiplication operation consists of a special operation for multiplying the packed weight value with the normal input value. And each operation was then processed in parallel using a GPU accelerator. As a result of this experiment, memory saved about 16 times than 32-bit DNN Model. Nevertheless, the accuracy was within 1%, similar to the 32-bit model.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.3
• /
• pp.613-623
• /
• 2011

In this paper, we present a method on the construction of quinternary logic circuits using Perfect shuffle. First, we discussed the input-output interconnection of quinternary logic function using Perfect Shuffle techniques and Kronecker product, and designed the basic cells of performing the transform matrix and the reverse transform matrix of quinternary Reed-Muller expansions(QRME) using addition circuit and multiplication circuit of GF(5). Using these basic cells and the input-output interconnection technique based on Perfect Shuffle and Kronecker product, we implemented the quinternary logic circuit based on QRME. The proposed design method of QRME is simple and very efficient to reduce addition circuits and multiplication circuits as compared with other methods for same logic function because of using matrix transform based on modular structures. The proposed design method of quinternary logic circuits is simple and regular for wire routing and possess the properties of concurrency and modularity of array.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.1
• /
• pp.123-136
• /
• 2004

In this paper, complexity and regularity of polynomial multiplication over $GF({2^n})$ are defined by using Hamming weight of rows and columns of the matrix ever GF(2) which represents polynomial multiplication. It is shown experimentally that in order to construct the block cipher robust against differential cryptanalysis, polynomial multiplication of substitution layer and the permutation layer should have high complexity and high regularity. With result of the experiment, a way of constituting S box and G function is suggested in the block cipher whose structure is similar to SEED, which is KOREA standard of 128-bit block cipher. S box can be formed with a nonlinear function and an affine transform. Nonlinear function must be strong with differential attack and linear attack, and it consists of an inverse number over $GF({2^8})$ which has neither a fixed pout, whose input and output are the same except 0 and 1, nor an opposite fixed number, whose output is one`s complement of the input. Affine transform can be constituted so that the input/output correlation can be the lowest and there can be no fixed point or opposite fixed point. G function undergoes linear transform with 4 S-box outputs using the matrix of 4${\times}$4 over $GF({2^8})$. The components in the matrix of linear transformation have high complexity and high regularity. Furthermore, G function can be constituted so that MDS(Maximum Distance Separable) code can be formed, SAC(Strict Avalanche Criterion) can be met, and there can be no weak input where a fixed point an opposite fixed point, and output can be two`s complement of input. The primitive polynomials of nonlinear function affine transform and linear transformation are different each other. The S box and G function suggested in this paper can be used as a constituent of the block cipher with high security, in that they are strong with differential attack and linear attack with no weak input and they are excellent at diffusion.

• Journal of KSNVE
• /
• v.7 no.1
• /
• pp.143-152
• /
• 1997

This paper introduces experimental results of the ride qulaity characteristics of automotive seats fixed on the vibration table that is noving simultaneously to the three-axis in a similar way to the real running condition. Vibration experiment was carried out for five different automotive seats and four Korean individuals. The assessment of the ride quality characteristics for each seat and indiviual was made not only from the analysis of vibration measurements but also from the evaluation of weighied vibration signals, which were obtained using the frequency weighting function and the multiplication factor dependent on the position and axis of vibration exposure to wehole-body. The usefulness of those assessment results in analysis of the ride quality of seats is discussed and their limitation is also pointed out in this paper.

• Korean Journal of Mathematics
• /
• v.20 no.4
• /
• pp.371-379
• /
• 2012

Let D be an integral domain, $\bar{D}$ be the integral closure of D, * be a star operation of finite character on D, $*_w$ be the so-called $*_w$-operation on D induced by *, X be an indeterminate over D, $N_*=\{f{\in}D[X]{\mid}c(f)^*=D\}$, and $Kr(D,*)=\{0\}{\cup}\{\frac{f}{g}{\mid}0{\neq}f,\;g{\in}D[X]$ and there is an $0{\neq}h{\in}D[X]$ such that $(c(f)c(h))^*{\subseteq}(c(g)c(h))^*$}. In this paper, we show that D is a *-quasi-Pr$\ddot{u}$fer domain if and only if $\bar{D}[X]_{N_*}=Kr(D,*_w)$. As a corollary, we recover Fontana-Jara-Santos's result that D is a Pr$\ddot{u}$fer *-multiplication domain if and only if $D[X]_{N_*} = Kr(D,*_w)$.

• Research in Mathematical Education
• /
• v.16 no.4
• /
• pp.245-263
• /
• 2012

The Korean mathematical curriculum has been changed three times during the resent five years. It led to changes in textbook system. In the 2007 revised mathematics curriculum, workbook was developed focusing on student's self-oriented learning, effective practice in differentiated classroom, and mathematics problem solving considering individual difference. This paper examines the characteristics of the tasks and the way the tasks are organized in the textbooks and the workbook in accordance with the 2007 revised mathematics curriculum; comparing with the function section before and after the amendment. Researchers examine whether the textbook and workbook were accomplished the purpose with "cognitive expectation", "level of cognitive demand", "and "response types". Researchers revised framework of [Son, J. W. & Senk, S. (2010). How reform curricula in the USA and Korea present multiplication and division of fraction. Educ. Stud. Math 74(2), 117-142] to make them suitable for the function section at the seventh grade.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.7
• /
• pp.1033-1041
• /
• 1990

In this paper, under pinch-off conditions, the gate-drain breakdown voltage characteristics of GaAs Power MESFET's as a function of device parameters such as channel thickness, doping concentration, gate length etc. are analyzed. Using the Green's function, the gate ionic charge induced by the depleted channel ionic charge is calculated. The impact ionization integral by avalanche multiplication between gate and drain is used to investigate breakdown phenomena. Especially, the localized excess surface charge effect as well as the uniform surface charge effect on breakdown voltage is considered.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.33 no.10
• /
• pp.381-386
• /
• 1984

This paper deals with the microcomputer realization of EMG pattern recognition system which provides identification of motion commands from the EMG signals for the on-line control of a prosthetic arm. A probabilistic model of pattern is formulated in the feature space of integral absolute value(IAV) to describe the relation between a motion command and the location of corresponding pattern. This model enables the derivation of sample density function of a command in the feature space of IAV. Classification is caried out through the multiclass sequential decision process, where the decision rule and the stopping rule of the process are designed by using the simple mathematical formulas defined as the likelihood probability and the decision measure, respectively. Some floating point algorithms such as addition, multiplication, division, square root and exponential function are developed for calculating the probability density functions and the decision measure. Only six primitive motions and one no motion are incorporated in this paper.

• Proceedings of the KIEE Conference
• /
• 1994.11a
• /
• pp.326-328
• /
• 1994

A modified Kohonen's simple Competitive Learning(SCL) algorithm which has binary reinforcement function and a constant adaptation gain is proposed. In contrast to the time-varing adaptation gain of the original Kohonen's SCL algorithm, the proposed algorithm uses a constant adaptation gain, and adds a binary reinforcement function in order to compensate for the lowered learning ability of SCL due to the constant adaptation gain. Since the proposed algorithm does not have the complicated multiplication, it's digital hardware implementation is much easier than one of the original SCL.

• The Journal of Information Technology
• /
• v.9 no.4
• /
• pp.49-55
• /
• 2006

In this paper, a new fragile watermarking algorithm for digital image is presented, which makes resolving the security and forgery problem of the digital image to be possible. The most suitable watermarking method that verifies the authentication and integrity of the digital image is the Wong's method, which invokes the hash function (MD5). The algorithm is safe because this method uses the hash function of the cryptology. The operations such as modulus, complement, shift, bitwise exclusive-or, bitwise inclusive-or are necessary for calculating the value of hash function. But, in this paper, an Arithmetic encoding method that only includes the multiplication operation is adopted. This technique prints out accumulative probability interval, which is obtained by multiplying the input symbol probability interval. In this paper, the initial probability interval is determined according to the value of the key, and the input sequence of the symbols is adjusted according to the key value so that the accumulative probability interval will depend on the key value. The integrity of the algorithm has been verified by experiment. The PSNR is above the 51.13db and the verifying time is $1/3{\sim}1/4$ of the verifying time of using the hash function (MD5), so, it can be used in the real-time system.