• Title/Summary/Keyword: pipeline model

Search Result 402, Processing Time 0.024 seconds

Development of an Automated ESG Document Review System using Ensemble-Based OCR and RAG Technologies

  • Eun-Sil Choi
    • Journal of the Korea Society of Computer and Information
    • /
    • v.29 no.9
    • /
    • pp.25-37
    • /
    • 2024
  • This study proposes a novel automation system that integrates Optical Character Recognition (OCR) and Retrieval-Augmented Generation (RAG) technologies to enhance the efficiency of the ESG (Environmental, Social, and Governance) document review process. The proposed system improves text recognition accuracy by applying an ensemble model-based image preprocessing algorithm and hybrid information extraction models in the OCR process. Additionally, the RAG pipeline optimizes information retrieval and answer generation reliability through the implementation of layout analysis algorithms, re-ranking algorithms, and ensemble retrievers. The system's performance was evaluated using certificate images from online portals and corporate internal regulations obtained from various sources, such as the company's websites. The results demonstrated an accuracy of 93.8% for certification reviews and 92.2% for company regulations reviews, indicating that the proposed system effectively supports human evaluators in the ESG assessment process.

A Pipelined Hash Join Method for Load Balancing (부하 균형 유지를 고려한 파이프라인 해시 조인 방법)

  • Moon, Jin-Gue;Park, No-Sang;Kim, Pyeong-Jung;Jin, Seong-Il
    • The KIPS Transactions:PartD
    • /
    • v.9D no.5
    • /
    • pp.755-768
    • /
    • 2002
  • We investigate the effect of the data skew of join attributes on the performance of a pipelined multi-way hash join method, and propose two new hash join methods with load balancing capabilities. The first proposed method allocates buckets statically by round-robin fashion, and the second one allocates buckets adaptively via a frequency distribution. Using hash-based joins, multiple joins can be pipelined so that the early results from a join, before the whole join is completed, are sent to the next join processing without staying on disks. Unless the pipelining execution of multiple hash joins includes some load balancing mechanisms, the skew effect can severely deteriorate system performance. In this paper, we derive an execution model of the pipeline segment and a cost model, and develop a simulator for the study. As shown by our simulation with a wide range of parameters, join selectivities and sizes of relations deteriorate the system performance as the degree of data skew is larger. But the proposed method using a large number of buckets and a tuning technique can offer substantial robustness against a wide range of skew conditions.

An FPGA Implementation of the Synthesis Filter for MPEG-1 Audio Layer III by a Distributed Arithmetic Lookup Table (분산산술연산방식을 이용한 MPEG-1 오디오 계층 3 합성필터의 FPGA 군현)

  • Koh Sung-Shik;Choi Hyun-Yong;Kim Jong-Bin;Ku Dae-Sung
    • The Journal of the Acoustical Society of Korea
    • /
    • v.23 no.8
    • /
    • pp.554-561
    • /
    • 2004
  • As the technologies of semiconductor and multimedia communication have been improved. the high-quality video and the multi-channel audio have been highlighted. MPEG Audio Layer 3 decoder has been implemented as a Processor using a standard. Since the synthesis filter of MPEG-1 Audio Layer 3 decoder requires the most outstanding operation in the entire decoder. the synthesis filter that can reduce the amount of operation is needed for the design of the high-speed processor. Therefore, in this paper, the synthesis filter. the most important part of MPEG Audio, is materialized in FPGA using the method of DAULT (distributed arithemetic look-up table). For the design of high-speed synthesis filter, the DAULT method is used instead of a multiplier and a Pipeline structure is used. The Performance improvement by 30% is obtained by additionally making the result of multiplication of data with cosine function into the table. All hardware design of this Paper are described using VHDL (VHIC Hardware Description Language) Active-HDL 6.1 of ALDEC is used for VHDL simulation and Synplify Pro 7.2V is used for Model-sim and synthesis. The corresponding library is materialized by XC4013E and XC4020EX. XC4052XL of XILINX and XACT M1.4 is used for P&R tool. The materialized processor operates from 20MHz to 70MHz.

A Load Balancing Method using Partition Tuning for Pipelined Multi-way Hash Join (다중 해시 조인의 파이프라인 처리에서 분할 조율을 통한 부하 균형 유지 방법)

  • Mun, Jin-Gyu;Jin, Seong-Il;Jo, Seong-Hyeon
    • Journal of KIISE:Databases
    • /
    • v.29 no.3
    • /
    • pp.180-192
    • /
    • 2002
  • We investigate the effect of the data skew of join attributes on the performance of a pipelined multi-way hash join method, and propose two new harsh join methods in the shared-nothing multiprocessor environment. The first proposed method allocates buckets statically by round-robin fashion, and the second one allocates buckets dynamically via a frequency distribution. Using harsh-based joins, multiple joins can be pipelined to that the early results from a join, before the whole join is completed, are sent to the next join processing without staying in disks. Shared nothing multiprocessor architecture is known to be more scalable to support very large databases. However, this hardware structure is very sensitive to the data skew. Unless the pipelining execution of multiple hash joins includes some dynamic load balancing mechanism, the skew effect can severely deteriorate the system performance. In this parer, we derive an execution model of the pipeline segment and a cost model, and develop a simulator for the study. As shown by our simulation with a wide range of parameters, join selectivities and sizes of relations deteriorate the system performance as the degree of data skew is larger. But the proposed method using a large number of buckets and a tuning technique can offer substantial robustness against a wide range of skew conditions.

Estimation of Applicability of Empirical Design Procedure for Predicting Seismic Response of Buried Gas Pipelines through 3D Time-history Analysis (3차원 시간이력해석을 통한 매설가스배관 종방향 지진응답 예측을 위한 경험적 설계법의 적용성 평가)

  • Kwak, Hyungjoo;Park, Duhee;Lee, Jangguen;Kang, Jaemo
    • Journal of the Korean Geotechnical Society
    • /
    • v.31 no.9
    • /
    • pp.53-68
    • /
    • 2015
  • Longitudinal strain is an important component of seismic design for buried pipelines. A design procedure which determines the wavelength from site natural period and shear wave velocity of the soil layer and closed-form solutions of pipelines under a harmonic motion is typically used in design. However, the applicability of the procedure has not yet been thoroughly investigated. In this paper, displacement-time histories extracted from 1D site response analyses are used in 3D shell-spring model to accurately predict the response of pipelines. The results are closely compared to those from the design procedure. The area of interest is East Siberia. Performing a site response analysis to determine site specific displacement time history is highlighted. The site natural period may be used to predict the predominant period of the acceleration time history, but cannot be used to estimate the predominant period of the displacement time history. If an accurate estimate of the predominant period of the displacement time history is provided, it is demonstrated that the design equation can be successfully used to predict the response of pipelines.

Necessity of the Physical Distribution Cooperation to Enhance Competitive Capabilities of Healthcare SCM -Bigdata Business Model's Viewpoint- (의료 SCM 경쟁역량 강화를 위한 물류공동화 도입 필요성 -빅데이터 비즈니스 모델 관점-)

  • Park, Kwang-O;Jung, Dae-Hyun;Kwon, Sang-Min
    • Management & Information Systems Review
    • /
    • v.39 no.3
    • /
    • pp.17-35
    • /
    • 2020
  • The purpose of this study is to develop business models for current situational scenarios reflecting customer needs emphasize the need for implementing a logistics cooperation system by analyzing big data to strengthen SCM competitiveness capacities. For healthcare SCM competitiveness needed for the logistics cooperation usage intent, they were divided into product quality, price leadership, hand-over speed, and process flexibility for examination. The wordcloud results that analyzed major considerations to realize work efficiency between medical institutes, words like unexpected situations, information sharing, delivery, real-time, delivery, convenience, etc. were mentioned frequently. It can be analyzed as expressing the need to construct a system that can immediately respond to emergency situations on the weekends. Furthermore, in addition to pursuing communication and convenience, the importance of real-time information sharing that can share to the efficiency of inventory management were evident. Accordingly, it is judged that it is necessary to aim for a business model that can enhance visibility of the logistics pipeline in real-time using big data analysis on site. By analyzing the effects of the adaptability of a supply chain network for healthcare SCM competitiveness, it was revealed that obtaining competitive capacities is possible through the implementation of logistics cooperation. Stronger partnerships such as logistics cooperation will lead to SCM competitive capacities. It will be necessary to strengthen SCM competitiveness by searching for a strategic approach among companies in a direction that can promote mutual partnerships among companies using the joint logistics system of medical institutes. In particular, it will be necessary to search for ways to utilize HCSM through big data analysis according to the construction of a logistics cooperation system.

Multi-Channel Pipelining for Energy Efficiency and Delay Reduction in Wireless Sensor Network (무선 센서 네트워크에서 에너지 효율성과 지연 감소를 위한 다중 채널 파리프라인 기법)

  • Lee, Yoh-Han;Kim, Daeyoung
    • Journal of the Institute of Electronics and Information Engineers
    • /
    • v.51 no.11
    • /
    • pp.11-18
    • /
    • 2014
  • Most of the energy efficient MAC protocols for wireless sensor networks (WSNs) are based on duty cycling in a single channel and show competitive performances in a small number of traffic flows; however, under concurrent multiple flows, they result in significant performance degradation due to contention and collision. We propose a multi-channel pipelining (MCP) method for convergecast WSN in order to address these problems. In MCP, a staggered dynamic phase shift (SDPS) algorithms devised to minimize end-to-end latency by dynamically staggering wake-up schedule of nodes on a multi-hop path. Also, a phase-locking identification (PLI) algorithm is proposed to optimize energy efficiency. Based on these algorithms, multiple flows can be dynamically pipelined in one of multiple channels and successively handled by sink switched to each channel. We present an analytical model to compute the duty cycle and the latency of MCP and validate the model by simulation. Simulation evaluation shows that our proposal is superior to existing protocols: X-MAC and DPS-MAC in terms of duty cycle, end-to-end latency, delivery ratio, and aggregate throughput.

A Study on Production and Experience of Immersive Contents based on Mixed Reality and Virtual Reality using Meta Quest Pro (메타 퀘스트 프로를 활용한 혼합현실과 가상현실 기반의 몰입형 콘텐츠 제작 및 경험에 관한 연구)

  • Jongseon Kim;Sumin Kong;Moonsu Jang;Jinmo Kim
    • Journal of the Korea Computer Graphics Society
    • /
    • v.30 no.3
    • /
    • pp.71-79
    • /
    • 2024
  • This study organizes an immersive content production pipeline using Meta Quest Pro as an asymmetric virtual environment where mixed reality(MR) and virtual reality(VR) users participate and interact together. Based on this, we compare and analyze the differences in presence and experience provided by the user's experience environment. The proposed production process is to build an integrated development environment using Meta XR All-in-One SDK based on the Unity 3D engine. This includes a real space analysis method using the Room Model function for organic and accurate interaction between MR users in the real world and VR users based on virtual scenes at fixed coordinates. Based on this, this study produces immersive table tennis content where MR and VR users participate together. Finally, we conduct survey experiments to compare and analyze the effects of differences in platform and participation methods on presence and experience using the produced content. As a result, this study confirmed that all users can feel a satisfactory presence and experience within an experience environment where real and virtual correspond.

Investigating Data Preprocessing Algorithms of a Deep Learning Postprocessing Model for the Improvement of Sub-Seasonal to Seasonal Climate Predictions (계절내-계절 기후예측의 딥러닝 기반 후보정을 위한 입력자료 전처리 기법 평가)

  • Uran Chung;Jinyoung Rhee;Miae Kim;Soo-Jin Sohn
    • Korean Journal of Agricultural and Forest Meteorology
    • /
    • v.25 no.2
    • /
    • pp.80-98
    • /
    • 2023
  • This study explores the effectiveness of various data preprocessing algorithms for improving subseasonal to seasonal (S2S) climate predictions from six climate forecast models and their Multi-Model Ensemble (MME) using a deep learning-based postprocessing model. A pipeline of data transformation algorithms was constructed to convert raw S2S prediction data into the training data processed with several statistical distribution. A dimensionality reduction algorithm for selecting features through rankings of correlation coefficients between the observed and the input data. The training model in the study was designed with TimeDistributed wrapper applied to all convolutional layers of U-Net: The TimeDistributed wrapper allows a U-Net convolutional layer to be directly applied to 5-dimensional time series data while maintaining the time axis of data, but every input should be at least 3D in U-Net. We found that Robust and Standard transformation algorithms are most suitable for improving S2S predictions. The dimensionality reduction based on feature selections did not significantly improve predictions of daily precipitation for six climate models and even worsened predictions of daily maximum and minimum temperatures. While deep learning-based postprocessing was also improved MME S2S precipitation predictions, it did not have a significant effect on temperature predictions, particularly for the lead time of weeks 1 and 2. Further research is needed to develop an optimal deep learning model for improving S2S temperature predictions by testing various models and parameters.

Design of Deep Learning-based Tourism Recommendation System Based on Perceived Value and Behavior in Intelligent Cloud Environment (지능형 클라우드 환경에서 지각된 가치 및 행동의도를 적용한 딥러닝 기반의 관광추천시스템 설계)

  • Moon, Seok-Jae;Yoo, Kyoung-Mi
    • Journal of the Korean Applied Science and Technology
    • /
    • v.37 no.3
    • /
    • pp.473-483
    • /
    • 2020
  • This paper proposes a tourism recommendation system in intelligent cloud environment using information of tourist behavior applied with perceived value. This proposed system applied tourist information and empirical analysis information that reflected the perceptual value of tourists in their behavior to the tourism recommendation system using wide and deep learning technology. This proposal system was applied to the tourism recommendation system by collecting and analyzing various tourist information that can be collected and analyzing the values that tourists were usually aware of and the intentions of people's behavior. It provides empirical information by analyzing and mapping the association of tourism information, perceived value and behavior to tourism platforms in various fields that have been used. In addition, the tourism recommendation system using wide and deep learning technology, which can achieve both memorization and generalization in one model by learning linear model components and neural only components together, and the method of pipeline operation was presented. As a result of applying wide and deep learning model, the recommendation system presented in this paper showed that the app subscription rate on the visiting page of the tourism-related app store increased by 3.9% compared to the control group, and the other 1% group applied a model using only the same variables and only the deep side of the neural network structure, resulting in a 1% increase in subscription rate compared to the model using only the deep side. In addition, by measuring the area (AUC) below the receiver operating characteristic curve for the dataset, offline AUC was also derived that the wide-and-deep learning model was somewhat higher, but more influential in online traffic.