SKSeong-Hwan Kim

Seong-Hwan Kim

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Seong-Hwan Kim has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • SLO-Aware DL Job Scheduling for Efficient FPGA-GPU Edge Cloud Computing

    by Taewoo Kim, Minsu Jeon, Changha Lee, SeongHwan Kim, Fawaz Al-Hazemi, Chan-Hyun Youn


  • Open Access

    Recursive Visual Explanations Mediation Scheme Based on DropAttention Model With Multiple Episodes Pool

    by Minsu Jeon, Taewoo Kim, Seonghwan Kim, Changha Lee, Chan-Hyun Youn.

    In some DL applications such as remote sensing, it is hard to obtain the high task performance ( e.g . accuracy) using the DL model on image analysis due to the low resolution characteristics of the imagery. Accordingly, several studies attempted to provide visual explanations or apply the attention mechanism to enhance the reliability on the image analysis. However, there still remains structural complexity on obtaining a sophisticated visual explanation with such existing methods: 1) which layer will the visual explanation be extracted from, and 2) which layers the attention modules will be applied to. 3) Subsequently, in order to observe the aspects of visual explanations on such diverse episodes of applying attention modules individually, training cost inefficiency inevitably arises as it requires training the multiple models one by one in the conventional methods. In order to solve the problems, we propose a new scheme of mediating the visual explanations in a pixel-level recursively. Specifically, we propose DropAtt that generates multiple episodes pool by training only a single network once as an amortized model, which also shows stability on task performance regardless of layer-wise attention policy. From the multiple episodes pool generated by DropAtt, by quantitatively evaluating the explainability of each visual explanation and expanding the parts of explanations with high explainability recursively, our visual explanations mediatio scheme attempts to adjust how much to reflect each episodic layer-wise explanation for enforcing a dominant explainability of each candidate. On the empirical evaluation, our methods show their feasibility on enhancing the visual explainability by reducing average drop about 17% and enhancing the rate of increase in confidence 3%.


  • Open Access

    FleX: A Flex Interconnected HPC System With Stochastic Load Balancing Scheme

    by Minsu Jeon, Kyung-No Joo, Taewoo Kim, Seonghwan Kim, Chan-Hyun Youn

    We propose a new low-diameter interconnection network called FleX, which offers high flexibility when installing interconnections in a HPC system. FleX consists of multiple layers with only connections between neighboring layers and not within each layer. These structural properties make it easy to achieve a low diameter with regardless of the scale. The cross-like connections between the adjacent layers in FleX impart various alternative minimal paths, allowing FleX to have high resiliency and a wide bisection width. We also discuss the minimal routing scheme and a stochastic load balancing scheme (LBR) for the proposed interconnection network. Through cycle-based simulations, the performance of FleX is evaluated, and the cost and power consumption analyses in comparison with other interconnection networks are also conducted. We verify that FleX has high configuration flexibility with regard to cost and performance, and also provides low latency and high saturation throughput with the same cost over the legacy interconnection networks for the HPC system. Moreover, being synergied with the proposing LBR, we also verify that FleX can expand its saturation throughput further while only sacrificing the latency slightly.

  • Pattern-Wise Embedding System for Scalable Time-series Database

    by Changha Lee, Seong-Hwan Kim, Chan-Hyun Youn

    The neural network has proved to solve a wide range of problem. However, it is difficult to adjust the model parameter in a time-varying natural language change. Therefore, this paper proposes the pattern-wise embedding system to address this problem. The proposed approach adopts the online learning with general representation and the temporary online forecasting. To obtain proper time-invariant information, the embedding vector is extracted. Using random sampled embedding vectors, the proposed framework adapts general representations. The proposed system applies temporary weight to increase performance of the nearest task. Experiments on real-world datasets show the effectiveness of our proposed framework.

  • Open Access

    Cooperating Edge Cloud-Based Hybrid Online Learning for Accelerated Energy Data Stream Processing in Load Forecasting

    by Changha Lee, Seong-Hwan Kim, Chan-Hyun Youn

    The data analysis platform used in smart grid is important to provide more accurate data validation and advanced power services. Recently, the researches based on deep neural network have been increasing in data analytic platforms to address various problems using artificial intelligence. The main problem to analyze multiple meter data based on deep learning is that the data distribution is varying according to both different client and time flow. Some studies, such as continual learning, are effective in dynamically fluctuating data distributions, but require additional complex computational procedures that make it difficult to construct an online learning system for processing data streams. In this paper, we proposed a hybrid deep learning scheduling algorithm to improve accuracy and accelerate learning performance in a multiple smart meter source environment, of which biased data feature varies dynamically. We use a simple analysis method, cosine similarity, to reduce computation complexity. By analyzing the frequency distribution of cosine similarity, a model recognizes that biased data feature of power consumption patterns. The skewed data distribution is reduced by using the zero skewness property of of an uniform distribution. The diversity of memory buffer was increased by update strategy which maximizes variance of pattern. When scheduling an online and offline gradient in different computational complexity, the proposed model reduces processing time by selectively calculating gradient considering the degree of data feature transition. To verify the performance of the proposed algorithm, we conducted three experiments with AMI stream data on the proposed method and the existing method of online learning. The experimental results demonstrate that our method can achieve reasonable performance in terms of trade-off between accuracy and processing time.


  • An Accelerated Continual Learning with Demand Prediction based Scheduling in Edge-Cloud Computing

    by Changha Lee, Seong-Hwan Kim, Chan-Hyun Youn

    As the development of smart grid with Advanced Metering Infrastructure (AMI) consisting of network infrastructure, smart meter, and data management system, the smart grid system can analyze energy data to efficiently control energy generation and distribution. Through recent advance of analysis based on neural network, some deep neural networks have proven to perform better than conventional analytical techniques. However, Basic learning process is facing challenges on analyze time-series data from smart meter based on deep learning in realtime. Although the strategies of gradually learning a deep neural network through the continual learning method was proposed, it is only effective when data feature is not significantly changed, therefore, the performance improvements are still needed on environment where the data distribution fluctuates according to different power consumption habits. Therefore, we proposed a scheduled continual deep learning on edge-cloud system to improve and accelerate learning performance on the multi-client power consumption data, which biased data feature varies dramatically. Using cosine similarity of electric load pattern, the scheduling algorithm manages and controls the gradient from optimizing process. The evaluated performance with general experiments shows the validity of proposed scheme compared to the base method.


  • Open Access

    An Accelerated Edge Cloud System for Energy Data Stream Processing Based on Adaptive Incremental Deep Learning Scheme

    by Seong-Hwan Kim, Changha Lee, Chan-Hyun Youn

    As smart metering technology evolves, power suppliers can make low-cost, low-risk estimation of customer-side power consumption by analyzing energy demand data collected in real-time. With advances network infrastructure, smart sensors, and various monitoring technologies, a standardized energy metering infrastructure, called advanced metering infrastructure (AMI), has been introduced and deployed to urban households to allow them to develop efficient power generation plans. Compared to traditional stochastic approaches for time-series data analysis, deep-learning methods have shown superior accuracy on many prediction applications. Because smart meters and infrastructure monitors produce a series of measurements over time, a large amount of data is accumulated, creating a large data stream, which takes much time from data generation to deployment of deep-learning model training. In this article, we propose an accelerated computing system that considers time-variant properties for accurate prediction of energy demand by processing the AMI stream data. The proposed system is a real-time training/inference system that deploys AMI data over a distributed edge cloud. It comprises two core components: an adaptive incremental learning solver and deep-learning acceleration with FPGA-GPU resource scheduling. An adaptive incremental learning scheme adjusts the batch/epoch in training iteration to reduce the time delay of the latest trained model, while trying to prevent biased-training due to the sub-optimal optimizer of incremental learning. In addition, a resource scheduling scheme manages various accelerator resources for accelerated deep-learning processing while minimizing the computational cost. The experimental results demonstrated that our method achieved good performance for adaptive batch size and epoch for incremental learning while guaranteeing a low inference error, a high model score, and queue stability with cost efficient processing.


  • LPC $_\mathrm{FreqSchd}$ FreqSchd : A local power controller using the frequency scheduling approach for virtualized servers

    by Fawaz Al-Hazemi, Dong-Ki Kang, Seong-Hwan Kim, Yuyang Peng, S. H. Shah Newaz, Chan-Hyun Youn


  • Adaptive VM Management with Two Phase Power Consumption Cost Models in Cloud Datacenter

    by Dong-Ki Kang, Fawaz Al-Hazemi, Seong-Hwan Kim, Min Chen, Limei Peng, Chan-Hyun Youn