• 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.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 1999.05a
• /
• pp.291-294
• /
• 1999

The effect of AI$_2$O$_3$+Y$_2$O$_3$additives on fracture toughness of $\beta$-SiC-ZrB$_2$composites by hot-pressed sintering were Investigated. The $\beta$-SiC-ZrB$_2$ ceramic composites were hot-presse sintered and annealed by adding 1, 2, 3wt% AI$_2$O$_3$+Y$_2$O$_3$(6:4wt%) powder as a liquid forming additives at 195$0^{\circ}C$ for 4h. In this microstructures, no reactions were observed between $\beta$-SiC and ZrB$_2$, and the relative density Is over 90.79% of the theoretical density and the porosity decreased with increasing AI$_2$O$_3$+Y$_2$O$_3$ contents. Owing to crack deflection and crack bridging of fracture toughness mechanism, the fracture toughness showed the highest of 5.5328MPa . m$^{1}$2/ for composites added with 2wt% AI$_2$O$_3$+Y$_2$O$_3$ additives at room temperature. But the standard deviation of fracture toughness of specimens decreased with increasing AI$_2$O$_3$+Y$_2$O$_3$ contents and showed the highest of 0.8624 for composite tilth 1wt%, AI$_2$O$_3$+Y$_2$O$_3$additives.

• Electronics and Telecommunications Trends
• /
• v.38 no.5
• /
• pp.1-11
• /
• 2023

The emergence of generative hyperscale artificial intelligence (AI) has enabled new services, such as image-generating AI and conversational AI based on large language models. Such services likely lead to the influx of numerous users, who cannot be handled using conventional AI models. Furthermore, the exponential increase in training data, computations, and high user demand of AI models has led to intensive hardware resource consumption, highlighting the need to develop domain-specific semiconductors for hyperscale AI. In this technical report, we describe development trends in technologies for hyperscale AI processors pursued by domestic and foreign semiconductor companies, such as NVIDIA, Graphcore, Tesla, Google, Meta, SAPEON, FuriosaAI, and Rebellions.

• Journal of Pharmacopuncture
• /
• v.13 no.3
• /
• pp.47-62
• /
• 2010

Objectives: Colitis is an inflammatory bowel disease characterized by colonic mucosal inflammation and chronic relapsing events represents. The purpose of this study is to evaluate effects of pharmacopuncture of anti-inflammatory herbal compound (AiC) applied to the different acupoints in the acute colitis induced by trinitrobenzenesulphonic acid (TNBS) intracolonic injection in rats. Methods: In Male Sprague - Dawley rats, weighing 250~400g, TNBS (5 mg/kg) was infused intrarectally through a silicon rubber catheter into the anus under isoflurane anaesthesia. Acupoints of LI4 (Hapkok), ST25 (Cheonchu), ST36 (Joksamni), and BL25 (Daejangsu) were intramuscularly injected by AiC, respectively (injection volume & times: 0.2 ml / acupoint, twice times on the 2nd & 3rd day). Expressions of cFos protein in the periaqueductal gray (PAG), locus coeruleus (LC), nucleus of solitary tract (Sol), and the 6th lumbar spinal cord (L6 s.c.) were observed at 24 hr after TNBS induced colitis by immunohistochemistry. Results: The expression of c-Fos protein in the L6 s.c., Sol, LC and PAG increased 24 hr after TNBS injection into colorectum as compared to normal and 50% ethanol treated group. AiC to LI4 inhibited the expression of c-Fos protein in Sol and PAG but not L6 s.c. and LC. AiC to ST36 showed significant inhibition the c-Fos expression in L6 s.c., Sol and PAG. AiC to ST25 only showed the effects in L6 s.c. and PAG. AiC to BL25 inhibited significantly the expression of c-Fos protein all over the areas. To investigate whether or not endogenous opioids are involved, intrathecal injection of naltrexone (30ug/30ul) was applied before the 2nd pharmacopuncture treatment 24 hr after TNBS-induced colitis in rat. Naltrexone reversed the inhibition of c-Fos protein expression in the spinal cord and brainstem. Conclusions: These data show that pharmacopuncture of Aic potently inhibits signal pathways ascending hypersensitivity of colorectum after TNBS induced colitis and depends on the endogenous opioids according to acupoints.

• Korean Journal of Microbiology
• /
• v.48 no.3
• /
• pp.207-211
• /
• 2012

Autoinduction is mediated by signaling molecules known as autoinducers (AIs) that are produced, released and detected by bacterium itself. We recently reported that Corynebacterium glutamicum possesses an autoinduction system which secretes autoinducers during the stationary-phase of growth, triggering the expression of acyltransferase gene. However, it is still not clear what may act as autoinducers for the autoinduction in C. glutamicum. In this study, we compared the inducing effects of cell-free culture fluids obtained from a number of microbes including Agrobacterium tumefaciens, Vibrio harveyi, and Escherichia coli. Fluids from A. tumefaciens did not increase the expression of acyltransferase, whereas fluids from V. harveyi BB120 ($AI-1^+$, $AI-2^+$) did. Interestingly, the expression was increased by the fluids obtained from the early exponential-phase culture of BB120. Furthermore, this induction was not observed by the fluids from autoinducer mutants of V. harveyi MM77 ($AI-1^-$, $AI-2^-$) and BB152 ($AI-1^-$, $AI-2^+$). Unlike the effect shown by BB152, fluids from E. coli ($AI-1^-$, $AI-2^+$) still induced the acyltransferase expression. Taken together, these results suggest that C. glutamicum autoinducers seem to be unidentified molecules which do not belong to AI-1 or AI-2.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.22 no.9
• /
• pp.717-723
• /
• 2009

Ni/Al ('/' denotes deposition sequence) contacts were deposited on Al-implanted 4H-SiC for ohmic contact formation, and the conduction properties were characterized and compared with those of Ni-only contacts. The thicknesses of the Ni and Al thin film were 30 nm and 300 nm, respectively, and the films were sequentially deposited bye-beam evaporation without vacuum breaking. Rapid thermal anneal (RTA) temperature was varied as follows : $840^{\circ}C$, $890^{\circ}C$, and $940^{\circ}C$. The specific contact resistivity of the Ni contact was about $^{\sim}2\;{\pm}\;10^{-2}\;{\Omega}{\cdot}cm^2$, However, with the addition of Al overlayer, the specific contact resistivity decreased to about $^{\sim}2\;{\pm}\;10^{-4}\;{\Omega}{\cdot}cm^2$, almost irrespective of RTA temperature. X-ray diffraction (XRD) analysis of the Ni contact confirmed the existence of various Ni silicide phases, while the results of Ni/Al contact samples revealed that Al-contaning phases such as $Al_3Ni$, $Al_3Ni_2$, $Al_4Ni_3$, and $Ab_{3.21}Si_{0.47}$ were additionally formed as well as the Ni silicide phases. Energy dispersive spectroscopy (EDS) spectrum showed interfacial reaction zone mainly consisting of Al and Si at the contact interface, and it was also shown that considerable amounts of Si and C have diffused toward the surface. This indicates that contact resistance lowering of the Ni/Al contacts is related with the formation of the formation of interfacial reaction zone containing Al and Si. From these results, possible mechanisms of contact resistance lowering by the addition of Al were discussed.

• Electronics and Telecommunications Trends
• /
• v.32 no.4
• /
• pp.1-10
• /
• 2017

4차 산업혁명의 핵심 엔진인 AI 기술이 인류의 희망과 우려 속에 전 지구적인 열풍에 휩싸여 있다. 이 와중에 거대 IT 업체들이 AI 기술 확보를 위한 치열한 경쟁에 나서서, 유망한 AI startup들의 인수합병에 몰두하고 있다. 또한, VC(Venture Capitalist)들도 AI startup에 대한 적극적인 투자에 나서고 있다. 본고에서는 이러한 글로벌 AI startup들에 대한 인수합병 현황 및 투자 트랜드를 살펴보고, AI startup 들이 집중하고 있는 연구개발 분야를 분석함으로써, 국내 AI 기술 경쟁력 확보를 위한 방향 설정에 도움을 줄 수 있도록 시사점을 제시 하고자한다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.14 no.9
• /
• pp.4394-4401
• /
• 2013

The purpose of this study was to investigate the association of blood pressures and serum lipids, obesity indices in adults who examined health checkup. Study subjects were 2,734 adults of 20 years and over (1,344 males, 1,390 females), who underwent health package check-up at an university hospital during one year of 2012. We observed the age, sex, SBP, DBP, TC, TG, HDL-C, LDL-C, AI, BMI, body fat rate and waist circumference of study subjects after confirmed IRB of the hospital. As a results, TC, TG, AI was significantly higher in the group of hypertension than that of normotensive in both sex, but in only female, HDL-C and LDL-C was higher in the each group of normotensive and hypertension. BMI, body fat rate and waist circumference was significantly higher in the group of hypertension than that of normotensive in both sex. SBP and DBP was positive correlated with TC, TG, AI, BMI, body fat rate and waist circumference, but HDL-C and LDL-C was each correlated with positive and negative in female only. Age-adjusted odds ratio for hypertension by blood serum lipids and obesity indices was significantly increased in the group of abnormal than normal of TC, TG, BMI, but HDL-C and waist circumference was each increased in the abnormal group of female, and in the abnormal group of male. Above results suggested that the blood pressure related with serum lipid levels and obesity indices.

• Electronics and Telecommunications Trends
• /
• v.35 no.5
• /
• pp.69-82
• /
• 2020

In many technologies, artificial intelligence (AI) is becoming an important topic for areas based on the field of big data. However, applied AI cases and the research status of radio wave technology are not widely known to the public. The spread of AI to other areas is being followed by radio wave technologies, and much effort is being taken to evolve it into intelligent radio wave technologies in the future. This paper presents the recent areas of interest in radio wave technology, such as spectral sharing, illegal spectrum monitoring, radar detection, radio wave medical imaging, and channel modeling; examines the requirements for applying AI; and describes the applied cases, research trends, and standardization efforts that apply AI technology to them. On this basis, we will discuss the prospects of AI application to the expected radio wave technology of the future.

• Electronics and Telecommunications Trends
• /
• v.37 no.6
• /
• pp.32-42
• /
• 2022

Recently, decentralized approaches to artificial intelligence (AI) development, such as federated learning are drawing attention as AI development's cost and time inefficiency increase due to explosive data growth and rapid environmental changes. Collaborative AI technology that dynamically organizes collaborative groups between different agents to share data, knowledge, and experience and uses distributed resources to derive enhanced knowledge and analysis models through collaborative learning to solve given problems is an alternative to centralized AI. This article investigates and analyzes recent technologies and applications applicable to the research of multi-agent collaboration of AI bots, which can provide collaborative AI functionality autonomously.