• Electronics and Telecommunications Trends
• /
• v.36 no.2
• /
• pp.32-42
• /
• 2021

The rapid growth of deep-learning applications has invoked the R&D of artificial intelligence (AI) processors. A dedicated software framework such as a compiler and runtime APIs is required to achieve maximum processor performance. There are various compilers and frameworks for AI training and inference. In this study, we present the features and characteristics of AI compilers, training frameworks, and inference engines. In addition, we focus on the internals of compiler frameworks, which are based on either basic linear algebra subprograms or intermediate representation. For an in-depth insight, we present the compiler infrastructure, internal components, and operation flow of ETRI's "AI-Ware." The software framework's significant role is evidenced from the optimized neural processing unit code produced by the compiler after various optimization passes, such as scheduling, architecture-considering optimization, schedule selection, and power optimization. We conclude the study with thoughts about the future of state-of-the-art AI compilers.

• Journal of Biomedical Engineering Research
• /
• v.27 no.5
• /
• pp.203-209
• /
• 2006

The aim of this study is to investigate the feasibility of ex vivo MR diffusion tensor imaging technique in order to observe the diffusion-contrast characteristics of human gastric tissues. On normal and pathologic gastric tissues, which have been fixed in a polycarbonate plastic tube filled with 10% formalin solution, laboratory made 3D diffusion tensor Turbo FLASH pulse sequence was used to obtain high resolution MR images with voxel size of $0.5{\times}0.5{\times}0.5mm^3\;using\;64{\times}32{\times}32mm^3$ field of view in conjunction with an acquisition matrix of $128{\times}64{\times}64$. Diffusion weighted- gradient pulses were employed with b values of 0 and $600s/mm^2$ in 6 orientations. The sequence was implemented on a clinical 3.0-T MRI scanner(Siemens, Erlangen, Germany) with a home-made quadrature-typed birdcage Tx/Rx rf coil for small specimen. Diffusion tensor values in each pixel were calculated using linear algebra and singular value decomposition(SVD) algorithm. Apparent diffusion coefficient(ADC) and fractional anisotropy(FA) map were also obtained from diffusion tensor data to compare pixel intensities between normal and abnormal gastric tissues. The processing software was developed by authors using Visual C++(Microsoft, WA, U.S.A.) and mathematical/statistical library of GNUwin32(Free Software Foundation). This study shows that 3D diffusion tensor Turbo FLASH sequence is useful to resolve fine micro-structures of gastric tissue and both ADC and FA values in normal gastric tissue are higher than those in abnormal tissue. Authors expect that this study also represents another possibility of gastric carcinoma detection by visualizing diffusion characteristics of proton spins in the gastric tissues.

• Journal of Korean Ophthalmic Optics Society
• /
• v.9 no.2
• /
• pp.481-489
• /
• 2004

Poly(triazine bissulfide)s were synthesized from 6-dibutylamino-1,3,5-triazine-2,4-dithiol with bis(4-chloro-3-nitrophenyl)sulfone in the presence of the phase transfer catalyst, and from m-dibromide xylene and p-dibromide xylene with 6-dibutylamino-1,3,5-triazine-2,4-dithiol in the presence of cetyltrimethyl ammonium bromide at $70^{\circ}C$ for 24h. We could acquire the good results about solubility, thermal property, and molecular weight to make cast film. These results are important as base for the synthesis of functionalization polymer material being optical plastic material. The maximum algebra viscosity was 0.57~1.40dl/g at the temperature more than $50{\sim}60^{\circ}C$.

• Journal of the Korean Mathematical Society
• /
• v.36 no.4
• /
• pp.787-811
• /
• 1999

Let M be the maximal ideal space of the Banach algebra $H^{\infty}$ of bounded analytic functions on the open unit disc $\triangle$. For a positive singular measure ${\mu}\;on\;{\partial\triangle},\;let\;{L_{+}}^1(\mu)$ be the set of measures v with $0\;{\leq}\;{\nu}\;{\ll}\;{\mu}\;and\;{{\psi}_{\nu}}$ the associated singular inner functions. Let $R(\mu)\;and\;R_0(\mu)$ be the union sets of $\{$\mid$\psiv$\mid$\;<\;1\}\;and\;\{$\mid${\psi}_{\nu}$\mid$\;<\;0\}\;in\;M\;{\setminus}\;{\triangle},\;{\nu}\;\in\;{L_{+}}^1(\mu)$, respectively. It is proved that if $S(\mu)\;=\;{\partial\triangle}$, where $S(\mu)$ is the closed support set of $\mu$, then $R(\mu)\;=\;R0(\mu)\;=\;M{\setminus}({\triangle}\;{\cup}\;M(L^{\infty}(\partial\triangle)))$ is generated by $H^{\infty}\;and\;\overline{\psi_{\nu}},\;{\nu}\;{\in}\;{L_1}^{+}(\mu)$. It is proved that %d{\theta}(S(\mu))\;=\;0$ if and only if there exists as Blaschke product b with zeros $\{Zn\}_n$ such that $R(\mu)\;{\subset}\;{$\mid$b$\mid$\;<\;1}\;and\;S(\mu)$ coincides with the set of cluster points of $\{Zn\}_n$. While, we proved that $\mu$ is a sum of finitely many point measure such that $R(\mu)\;{\subset}\;\{$\mid${\psi}_{\lambda}$\mid$\;<\;1}\;and\;S(\lambda)\;=\;S(\mu)$. Also it is studied conditions on \mu for which $R(\mu)\;=\;R0(\mu)$.

• Communications of Mathematical Education
• /
• v.23 no.4
• /
• pp.977-998
• /
• 2009

In this article, we give a comparative study on the last 300 years of USA and Korean tertiary mathematics. The first mathematics classes in United States were offered before July, 1638, but the real founding of tertiary mathematics courses was in 1640 when Henry Dunster assumed the duties of the presidency at Harvard. President Dunster read arithmetics and geometry on Mondays and Tuesdays to the third year students during the first three quarters, and astronomy in the last quarter. So tertiary mathematics education in United States began at Harvard which is the oldest college in USA. After 230 years since then, Benjamin Peirce in 1870 made a major and first American contribution to mathematics and got an attention from European mathematicians. Major change on the role of Harvard mathematics from teaching to research made by G.D. Birkhoff when he joined as an assistant professor in 1912. Tertiary mathematics education in Korea started long before Chosun Dynasty. But it was given to only small number of government actuarial officers. Modern mathematics education of tertiary level in Korea was given at Sungkyunkwan, Ewha, Paichai, and Soongsil. But all college level education opportunity, particularly in mathematics, was taken over by colonial government after 1920. And some technical and normal schools offered some tertiary mathematics courses. There was no college mathematics department in Korea until 1945. After the World War II, the first college mathematics department was established, and Rimhak Ree in 1949 made a major and first Korean contribution to modern mathematics, and later found Ree group. He got an attention from western mathematicians for the first time as a Korean. It can be compared with Benjamin Peirce's contribution for USA.

• Journal of Korean Society of Forest Science
• /
• v.90 no.4
• /
• pp.535-542
• /
• 2001

The objective of this study is to develop a spatial analysis modeling technique to evaluate the functional suitability of forest lands for land slide prevention. The functional suitability is classified into 3 categories of high, medium and low according to the potential of land slide on forest lands. The potential of land slide hazards is estimated using the measurements of 7 major site factors : slope, bed rock, soil depth, shape of slope, forest type and D.B.H. class of trees. The analytic hierarchical process is applied to determining the relative weight of site factors in estimating the potential of land slides. The spatial analysis modeling starts building base layers for the 7 major site factors by $25m{\times}25m$ grid analysis or TIN analysis, reclassifies them and produces new layers containing standardized attribute values, needed in estimating land slide potential. To these attributes, applied is the weight for the corresponding site factor to build the suitability classification map by map algebra analysis. Then, finally, cell-grouping operations convert the suitability classification map to the land unit function map. The whole procedures of the spatial analysis modeling are presented in this paper.

• Korean Journal of Childcare and Education
• /
• v.10 no.2
• /
• pp.175-192
• /
• 2014

The purpose of this study was to investigate the differences in perceived importance of mathematical education (perceived importance) and mathematical interaction of 5-year-olds' mothers according to contents of mathematical education. Second, we intended to examine whether mothers' understanding of purpose of mathematical education predicted on their perceived importance and mathematical interaction. Third, we analyzed relative influence between mothers' perceived effectiveness of concrete materials and worksheets on their perceived importance and mathematical interaction. The subjects consisted of 151 mothers of 5-year-olds lived in D city and K province in Korea. The results were as follows: First, mothers' perceived importance and mathematical interaction were higher in 'number and arithmetic'. Second, mothers' understanding of purpose of mathematical education predicted their perceived importance in all contents and mathematical interaction in 'number and arithmetic', 'geometry', and 'algebra'. Third, mothers' perceived effectiveness of concrete materials predicted better in most contents of mathematical education. Meanwhile, in 'number and operation', mothers' perceived effectiveness of worksheets did a predictive role in their importance awareness. These results were discussed in terms of necessity of a parent education program to provide practical information about contents and methods of mathematical education for their 5-year-old children.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.4
• /
• pp.35-41
• /
• 2001

This paper describes Linear Discriminant Analysis and common vector extraction for speech recognition. Voice signal contains psychological and physiological properties of the speaker as well as dialect differences, acoustical environment effects, and phase differences. For these reasons, the same word spelled out by different speakers can be very different heard. This property of speech signal make it very difficult to extract common properties in the same speech class (word or phoneme). Linear algebra method like BT (Karhunen-Loeve Transformation) is generally used for common properties extraction In the speech signals, but common vector extraction which is suggested by M. Bilginer et at. is used in this paper. The method of M. Bilginer et al. extracts the optimized common vector from the speech signals used for training. And it has 100％ recognition accuracy in the trained data which is used for common vector extraction. In spite of these characteristics, the method has some drawback-we cannot use numbers of speech signal for training and the discriminant information among common vectors is not defined. This paper suggests advanced method which can reduce error rate by maximizing the discriminant information among common vectors. And novel method to normalize the size of common vector also added. The result shows improved performance of algorithm and better recognition accuracy of 2％ than conventional method.