• Title/Summary/Keyword: loss function

Search Result 2,962, Processing Time 0.036 seconds

The Design of Iron Loss Minimization of 600W IPMSM by Quasi-newton Method (Quasi-Newton Method에 의한 600W IPMSM의 철손 최소화 설계)

  • Baek, Sung-min;Cho, Gyu-won;Kim, Gyu-tak
    • The Transactions of The Korean Institute of Electrical Engineers
    • /
    • v.66 no.7
    • /
    • pp.1053-1058
    • /
    • 2017
  • In this paper, the design of iron loss minimization of 600W was performed by using Quasi-Newton method. Stator shoe, the width of stator teeth and yoke, and the length of d-axis flux path were selected as design parameters, and the output characteristics according to each design variable were considered. The objective function was set to minimize iron loss. Using the Quasi-Newton method, the variables converged to the target value while changing simultaneously and multiple times. As the algorithm advanced optimization, the correlation with the behavior of each variable was compared and analyzed.

Analysis of heat exchanger in the drying system using solar collector with evacuated tubes (진공관형 태양열 집열기를 이용한 건조장치의 열교환기 해석)

  • Kang, Hyung-Suk;Han, Young-Min;Lee, Gwi-Hyun;Lee, Sung-Joo;Yoon, Sae-Chang
    • New & Renewable Energy
    • /
    • v.2 no.1 s.5
    • /
    • pp.46-55
    • /
    • 2006
  • The performance enhancement of heat exchanger in the drying system using solar collector with evacuated tubes is analyzed. First, for this analysis, the heat loss from a reversed trapezoidal fin attached at the pipe is calculated as a function of convection characteristic number ratio, fin base length and fin tip length. Also, the optimum heat loss and fin tip length of the fin under certain conditions are presented. The overall surface effectiveness of the cylinder with reversed trapezoidal fins in the heat exchanger are shown as a function of half fin base height, fin lateral slope and fin tip length.

  • PDF

Influence on overfitting and reliability due to change in training data

  • Kim, Sung-Hyeock;Oh, Sang-Jin;Yoon, Geun-Young;Jung, Yong-Gyu;Kang, Min-Soo
    • International Journal of Advanced Culture Technology
    • /
    • v.5 no.2
    • /
    • pp.82-89
    • /
    • 2017
  • The range of problems that can be handled by the activation of big data and the development of hardware has been rapidly expanded and machine learning such as deep learning has become a very versatile technology. In this paper, mnist data set is used as experimental data, and the Cross Entropy function is used as a loss model for evaluating the efficiency of machine learning, and the value of the loss function in the steepest descent method is We applied the GradientDescentOptimize algorithm to minimize and updated weight and bias via backpropagation. In this way we analyze optimal reliability value corresponding to the number of exercises and optimal reliability value without overfitting. And comparing the overfitting time according to the number of data changes based on the number of training times, when the training frequency was 1110 times, we obtained the result of 92%, which is the optimal reliability value without overfitting.

Geoacoustic Inversion via Transmission Loss Matching and Matched Field Processing (전달손실 비교를 통한 지음향학적 인자 역산과 정합장처리)

  • Kim Kyungseop;Park Cheolsoo;Kim Seongil;Seong Woojae
    • The Journal of the Acoustical Society of Korea
    • /
    • v.24 no.6
    • /
    • pp.325-333
    • /
    • 2005
  • This paper proposes a geoacoustic inversion method for the experimental data or MAPLE 2004 experiment conducted in the East Sea of Korea in 2004 and shows source tracking test results to validate the Proposed inversion method. An objective function is defined as a correlation function of the measured and the simulated transmission loss data. The measured transmission data were obtained using a multi-tonal towed source and VLA. The VFSA (Very Fast Simulated Annealing) is applied to the inversion Problem which optimizes the objective function. After performing the inversion process for the S frequencies tonal data independently. geoacoustic models are constructed. Finally matched-field source tracking is Performed using the inverted parameters to verify them.

A Study on the Cover Ratio and the Sizing System of Apparels for Obese Women (비만 여성의 의복 치수체계 및 커버율에 판한 연구)

  • 이진희
    • Journal of the Korean Society of Clothing and Textiles
    • /
    • v.22 no.6
    • /
    • pp.737-748
    • /
    • 1998
  • This study was carried out on 132 obese women who satisfied both of conditions for obesity: equal to or over 1.6 in Rohrer index and 90 in bust girth. The purpose of the study was to set up a sizing system using the loss function which would be a guide for obese women to select ready-to-wear of suitable size. The results were as follows. 1) In the sizing system for large size apparel industry, each company has 4 to 7 sizes that differ in their content and number. Producing only 5 sizes was trying to minimize the producti on expenses. 2) The sizing system according to the loss function was the follwings. The height was 3: 149, 154.5 and 161 cm. The bust girth was 5:96.5, 100.5, 104, 107.5, 112 cm. The hip girth was 5: 95, 99, 102.5, 105.5, 110 cm. 3) In comparing the cover ratio of the newly suggested sizing system for obese women's garment with that of the Korea Sizing system for women's garment, the former was founded to have the greater cover ratio.

  • PDF

Determination of Control Limits of Conditional Variance Investigation: Application of Taguchi's Quality Loss Concept (조건부 차이조사의 관리한계 결정: 다구찌 품질손실 개념의 응용)

  • Pai, Hoo Seok;Lim, Chae Kwan
    • Journal of Korean Society for Quality Management
    • /
    • v.49 no.4
    • /
    • pp.467-482
    • /
    • 2021
  • Purpose: The main theme of this study is to determine the optimal control limit of conditional variance investigation by mathematical approach. According to the determination approach of control limit presented in this study, it is possible with only one parameter to calculate the control limit necessary for budgeting control system or standard costing system, in which the limit could not be set in advance, that's why it has the advantage of high practical application. Methods: This study followed the analytical methodology in terms of the decision model of information economics, Bayesian probability theory and Taguchi's quality loss function concept. Results: The function suggested by this study is as follows; ${\delta}{\leq}\frac{3}{2}(k+1)+\frac{2}{\frac{3}{2}(k+1)+\sqrt{\{\frac{3}{2}(k+1)\}^2}+4$ Conclusion: The results of this study will be able to contribute not only in practice of variance investigation requiring in the standard costing and budgeting system, but also in all fields dealing with variance investigation differences, for example, intangible services quality control that are difficult to specify tolerances (control limit) unlike tangible product, and internal information system audits where materiality standards cannot be specified unlike external accounting audits.

Synthetic Infra-Red Image Dataset Generation by CycleGAN based on SSIM Loss Function (SSIM 목적 함수와 CycleGAN을 이용한 적외선 이미지 데이터셋 생성 기법 연구)

  • Lee, Sky;Leeghim, Henzeh
    • Journal of the Korea Institute of Military Science and Technology
    • /
    • v.25 no.5
    • /
    • pp.476-486
    • /
    • 2022
  • Synthetic dynamic infrared image generation from the given virtual environment is being the primary goal to simulate the output of the infra-red(IR) camera installed on a vehicle to evaluate the control algorithm for various search & reconnaissance missions. Due to the difficulty to obtain actual IR data in complex environments, Artificial intelligence(AI) has been used recently in the field of image data generation. In this paper, CycleGAN technique is applied to obtain a more realistic synthetic IR image. We added the Structural Similarity Index Measure(SSIM) loss function to the L1 loss function to generate a more realistic synthetic IR image when the CycleGAN image is generated. From the simulation, it is applicable to the guided-missile flight simulation tests by using the synthetic infrared image generated by the proposed technique.

Comparative Analysis of VT-ADL Model Performance Based on Variations in the Loss Function (Loss Function 변화에 따른 VT-ADL 모델 성능 비교 분석)

  • Namjung Kim;Changjoon Park;Junhwi Park;Jaehyun Lee;Jeonghwan Gwak
    • Proceedings of the Korean Society of Computer Information Conference
    • /
    • 2024.01a
    • /
    • pp.41-43
    • /
    • 2024
  • 본 연구에서는 Vision Transformer 기반의 Anomaly Detection and Localization (VT-ADL) 모델에 초점을 맞추고, 손실 함수의 변경이 MVTec 데이터셋에 대한 이상 검출 및 지역화 성능에 미치는 영향을 비교 분석한다. 기존의 손실 함수를 KL Divergence와 Log-Likelihood Loss의 조합인 VAE Loss로 대체하여, 성능 변화를 심층적으로 조사했다. 실험을 통해 VAE Loss로의 전환은 VT-ADL 모델의 이상 검출 능력을 현저히 향상시키며, 특히 PRO-score에서 기존 대비 약 5%의 개선을 보였다는 점을 확인하였다. 이러한 결과는 손실 함수의 최적화가 VT-ADL 모델의 전반적인 성능에 중요한 영향을 미칠 수 있음을 시사한다. 또한, 이 연구는 Vision Transformer 기반 모델의 이상 검출과 지역화 작업에 있어서 손실 함수 선택의 중요성을 강조하며, 향후 관련 연구에 유용한 기준을 제공할 수 있을 것으로 기대된다.

  • PDF

Familial Hearine Loss (가족성 난청)

  • 전광수;장선문;박찬일
    • Proceedings of the KOR-BRONCHOESO Conference
    • /
    • 1978.06a
    • /
    • pp.7.2-7
    • /
    • 1978
  • There are over 60 types of hereditary hearing loss in man and most of these are distinguished from one another by outstanding associated defects caused the same gene producing hearing loss. However, there are at least 12 types of hereditary hearing loss with no known associated defects. The authors experienced two similar families with hereditary hearing loss. The affected offsprings had inevitably an affected parent in type of transmissin (autosomal dominant) and the hearing loss developed in childhood and progressed. Audiologic findings showed bilateral symmetrical moderate to severe sensori-neural hearing loss with flattened configuration on the audiogram. Vestibular function tests showed no abnormal findings and the associated anomalies could'nt be observed in history or clinical examinations.

  • PDF