• ETRI Journal
• /
• v.45 no.2
• /
• pp.318-328
• /
• 2023

Recently, embedded systems, such as mobile platforms, have multiple processing units that can operate in parallel, such as centralized processing units (CPUs) and neural processing units (NPUs). We can use deep-learning compilers to generate machine code optimized for these embedded systems from a deep neural network (DNN). However, the deep-learning compilers proposed so far generate codes that sequentially execute DNN operators on a single processing unit or parallel codes for graphic processing units (GPUs). In this study, we propose PartitionTuner, an operator scheduler for deep-learning compilers that supports multiple heterogeneous PUs including CPUs and NPUs. PartitionTuner can generate an operator-scheduling plan that uses all available PUs simultaneously to minimize overall DNN inference time. Operator scheduling is based on the analysis of DNN architecture and the performance profiles of individual and group operators measured on heterogeneous processing units. By the experiments for seven DNNs, PartitionTuner generates scheduling plans that perform 5.03% better than a static type-based operator-scheduling technique for SqueezeNet. In addition, PartitionTuner outperforms recent profiling-based operator-scheduling techniques for ResNet50, ResNet18, and SqueezeNet by 7.18%, 5.36%, and 2.73%, respectively.

• Journal of Practical Engineering Education
• /
• v.15 no.3
• /
• pp.671-679
• /
• 2023

This study aims to establish an evaluation model by quantifying the evaluation index as a follow-up study to the development of evaluation index for work-study parallel learning companies. An evaluation model was established by verifying the 2nd level components based on the quantitative factors of the learning company, the qualitative factors, the competency factors of the person in charge, and the competency factors of the learning workers, which are the highest-level components derived from previous study. For the evaluation of a learning company, an AHP survey was conducted with experts in charge of the company consulting to derive important factors that determine the quality of on-site education and training, and the evaluation model of the learning company was completed and grouped by calculating the weight between evaluation items proceeded. Work-study parallel program was promoted as a key policy to resolve the mismatch between industrial sites and school education and realize a competency-centered society, and as of December 2022, 16,664 companies participated in the training. Learning companies play a very important role as education and training supply organizations that conduct field training. It is expected that the support and consulting plan for each level of learning companies according to the evaluation model presented in this study will be used as basic data to improve the quality of work-study parallel program.

• Journal of Practical Engineering Education
• /
• v.12 no.2
• /
• pp.331-337
• /
• 2020

This study investigated the perception of the learning-workers' organizational effectiveness. They participate in the degree-linked work-learning parallel system. As a result, the turnover intention and job performance were higher than those of general office workers, and organizational commitment was lower. In addition, the higher scholastic year and grades, the higher job satisfaction. the higher grades, the lower burnout. specially the 2nd, 3rd, and 4th graders showed higher turnover intentions than the first grader. The results of the survey of (dis)satisfaction factors of learning workers also show that OJT management in the field should be strengthened rather than OFF-JT education in universities. This paper suggests that companies participating in the work-learning parallel system need higher-level management in aspects such as educational supports, personnel management, and OJT management as well as competency development of learning-workers. It shows that supplementary measures are needed for learning time and learner management of participating companies.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.12
• /
• pp.69-75
• /
• 2021

Recently, deep learning, which is the most used in the field of artificial intelligence, has a structure that is gradually becoming larger and more complex. As the deep learning model grows, a large amount of data is required to learn it, but there are cases in which it is difficult to integrate and learn the data because the data is distributed among several owners and security issues. In that situation we conducted parallel learning for each users that own data and then studied how to integrate it. For this, distributed learning was performed for each owner assuming the security situation as V-environment and H-environment, and the results of distributed learning were integrated using Average, Max, and AbsMax. As a result of applying this to the mnist-fashion data, it was confirmed that there was no significant difference from the results obtained by integrating the data in the V-environment in terms of accuracy. In the H-environment, although there was a difference, meaningful results were obtained.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.10 no.4
• /
• pp.288-294
• /
• 2017

Recently, deep learning using convolutional neural network (CNN) has been extensively studied in image recognition. Convolution consists of addition and multiplication. Multiplication is computationally expensive in hardware implementation, relative to addition. It is also important factor limiting a chip design in an embedded deep learning system. In this paper, I propose a parallel-addition processing algorithm that converts grayscale images to the superposition of binary images and performs convolution only with addition. It is confirmed that the convolution can be performed by a parallel-addition method capable of reducing the processing time in experiment for verifying the availability of proposed algorithm.

• The Transactions of the Korean Institute of Electrical Engineers B
• /
• v.48 no.6
• /
• pp.294-300
• /
• 1999

There is a growing interest in hybrid electric vehicles due to environmental concerns. Recent efforts are directed toward developing an improved main component systems for the hybrid electric vehicle applications. Soon after the introduction of electric starter for internal combustion engine early this century, despite being energy efficient and nonpolluting, electric vehicle lost the battle completly to internal combustion engine due to its limited range and inferior performance. Hybrid Electric vehicles offer the most promising solutions to reduce the emission of vehicles. This paper describes a method for cost reduction estimation of parallel hybrid electric vehicle. We used a cost reduction structure that consisted of five major subsystems (three-type and two-type motor) for parallel hybrid electric vehicle. Especially, we estimated the potential for cost reductions in parallel hybrid electric vehicle as a function of time using the learning curve. Also, we estimated the potentials of cost by depreciation.

• International Journal of Advanced Culture Technology
• /
• v.12 no.2
• /
• pp.7-15
• /
• 2024

Despite a recent surge in multifaceted research on AI-integrated language learning, empirical studies in this area remain limited. This study adopts a Human-Generative AI parallel processing model to examine students' perceptions, asking 182 college students to independently construct knowledge and then compare their efforts with the results generated through in-classroom conversations with ChatGPT 3.5. In questionnaire responses, most students indicated that they found these activities useful and expressed a keen interest in learning various ways to utilize generative AI for language learning with instructor guidance. The findings confirm that ChatGPT's potential as a virtual conversation partner. Identifying specific reasons for the perceived usefulness of conversation activities and drawbacks of ChatGPT, this study emphasizes the importance of teachers staying informed about both the latest advances in technology and their limitations. We recommend that teachers endeavor to creatively design various classroom activities using AI technology.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.2
• /
• pp.205-211
• /
• 2024

The purpose of this study is to implement an agent that intelligently performs tracking and movement through reinforcement learning using the Unity and ML-Agents. In this study, we conducted an experiment to compare the learning performance between training one agent in a single learning simulation environment and parallel training of several agents simultaneously in a multi-learning simulation environment. From the experimental results, we could be confirmed that the parallel training method is about 4.9 times faster than the single training method in terms of learning speed, and more stable and effective learning occurs in terms of learning stability.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.47 no.2
• /
• pp.107-115
• /
• 2024

With the advent of the 4th Industrial Revolution, changes in the market environment and employment environment are accelerating due to smart technological innovation, and securing professional manpower and developing human resources for domestic small and medium-sized enterprises is becoming very important. Recently, most of the domestic small and medium-sized enterprises are experiencing hiring difficulties, and the development and training of human resources to overcome this is still lacking in systemization, despite much support from the government. This reflects the reality that it is not easy to invest training costs and time to adapt new employees to small and medium-sized businesses. Based on these problems, the work-study parallel project was introduced to cultivate practical talent in small and medium-sized businesses. Work-study parallel training is carried out in the form of mentoring between corporate field teachers and learning workers in actual workplaces, and even if the training is the same, there are differences depending on the learner's attitude, learning motivation, and training achievement. Ego state is a theory that can identify personality types and has the advantage of being able to understand and acknowledge oneself and others and intentionally improve positive factors to induce optimized interpersonal relationships. Accordingly, the purpose of this study is to analyze the attitudes of learning workers, who are the actual subjects for improving the performance of work-study parallel projects and establishing a stable settlement within the company, based on their ego status. Through this study, we aim to understand the impact of the personality type of learning workers on training performance and to suggest ways to improve training performance through work-study parallelism.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.33B no.2
• /
• pp.149-158
• /
• 1996

In this paper we describe designing and implementing a digital neural chip and a parallel neural machine for simulating large scale neural netsorks. The chip is a single-chip multiprocessor which has four digiral neural processors (DNP-II) of the same architecture. Each DNP-II has program memory and data memory, and the chip operates in MIMD (multi-instruction, multi-data) parallel processor. The DNP-II has the instruction set tailored to neural computation. Which can be sed to effectively simulate various neural network models including on-chip learning. The DNP-II facilitates four-way data-driven communication supporting the extensibility of parallel systems. The parallel neural machine consists of a host computer, processor boards, a buffer board and an interface board. Each processor board consists of 8*8 array of DNP-II(equivalently 2*2 neural chips). Each processor board acn be built including linear array, 2-D mesh and 2-D torus. This flexibility supports efficiency of mapping from neural network models into parallel strucgure. The neural system accomplishes the performance of maximum 40 GCPS(giga connection per second) with 16 processor boards.