• Title/Summary/Keyword: ensembles

Search Result 116, Processing Time 0.024 seconds

BERT-Based Logits Ensemble Model for Gender Bias and Hate Speech Detection

  • Sanggeon Yun;Seungshik Kang;Hyeokman Kim
    • Journal of Information Processing Systems
    • /
    • v.19 no.5
    • /
    • pp.641-651
    • /
    • 2023
  • Malicious hate speech and gender bias comments are common in online communities, causing social problems in our society. Gender bias and hate speech detection has been investigated. However, it is difficult because there are diverse ways to express them in words. To solve this problem, we attempted to detect malicious comments in a Korean hate speech dataset constructed in 2020. We explored bidirectional encoder representations from transformers (BERT)-based deep learning models utilizing hyperparameter tuning, data sampling, and logits ensembles with a label distribution. We evaluated our model in Kaggle competitions for gender bias, general bias, and hate speech detection. For gender bias detection, an F1-score of 0.7711 was achieved using an ensemble of the Soongsil-BERT and KcELECTRA models. The general bias task included the gender bias task, and the ensemble model achieved the best F1-score of 0.7166.

A Study on the Comfort and Skin Temperature on the Clothing Environment in the Taegu City during Years -Concentrated on indoor uniforms- (섬유·패션 산업 DB 구축을 위한 대구지역의 연간 피복환경내의 피부온과 쾌적성 -실내 제복을 중심으로-)

  • Ryo, Duck-Hwan;Lee, Uk-Ja;Kim, Seong-Jin;Song, Min-Kyo;Cho, Ji-Hyun;Jung, Meung-Sun
    • Fashion & Textile Research Journal
    • /
    • v.1 no.4
    • /
    • pp.376-386
    • /
    • 1999
  • The purpose of this study was to determine the physiological responses and the Subjective sensations of the human subjects when 10 human subjects(5 males and 5 females) were tested with selected ensembles, including the uniforms of students, industrial and bank workers. For the study, garment material and design were selected, which was mostly used in Taegu area by the survey. Thermal manikin and the human subject tests were performed. The results were as follows: 1. There was no significant difference among skin temperature of head by season and among that of breast and back by all variables, however, there was a significant difference among that of belly by gender and that of thigh and shank by season. In addition, there was no significant difference among rectal temperature by gender and season, but there was a significant difference by uniform. 2. Mean skin temperature of male subjects was significantly higher than that of female subjects. Specially in lower environmental condition, the difference got larger. 3. There was a significant difference among purse rate of subjects by uniform, but there was no significant difference among those by gender and season. There were no significant difference among maximum and minimum blood pressure by gender, uniform, and season. 4. There was a significant difference in temperature of chest among gender and uniform, and there was no significant difference in temperature of back by season. In addition, There was a significant difference in humidity of back by uniform and season but there were no significant difference in that by gender. 5. There was a significant difference in humidity sensation by gender, comfort sensation by uniform and season, but there was no significant difference thermal sensation by uniform and season. 6. For the result of regression analysis, we got the regression equations as follows: Clo=1.810 Thickness+0.525, Clo=0.475 weight+0.863.

  • PDF

Cocaine-induced Changes in Functional Connectivities between Simultaneously Recorded Single Neurons in the SI Cortex and the VPL Thalamus of Conscious Rats

  • Shin, Hyung-Cheul;Park, Hyoung-Jin;Oh, Yang-Seok;Chapin, John K.
    • The Korean Journal of Physiology
    • /
    • v.27 no.1
    • /
    • pp.79-91
    • /
    • 1993
  • The present study was carried out to determine the effects of cocaine (0.25, 1.0, 10.0 mg/kg, i.p.) on the interactions between spontaneously active neurons within ensembles of simultaneously recorded neurons in the primary somatosensory cortex (Sl, n= 20) and the ventroposterolateral (VPL, n= 16) thalamic nucleus of awake rats. Spike triggered cross correlation histograms were constructed between pairs of simultaneously recorded neurons. Among 101 neuronal pairs analyzed, 22.7% showed correlations indicative of various functional connections among the cortical cells, two corticothalamic interactions and one thalamocortical excitatory interaction. There were also 15 cofiring activities among SI cortical cells. These functional connectivities appeared to be modulated (weakened, abolished, or strengthened) during the 5 to 30 min following cocaine injection. The effects of saline were tested as a control, but it did not appear to alter the functional connectivities. In general, cocaine-induced changes of the functional interactions were mainly due to the concomitant alterations of the uncorrelated background discharges. These results suggest that the biphasic effects of cocaine on the spontaneously established neural networks among the SI cortical and the VPL thalamic cells of conscious rat were mainly indirect. However, various changes of the functional interactions by different doses of cocaine appeared to be a possible neural network mechanism for the cocaine induced modulation of afferent somatosensory transmission.

  • PDF

Conformational Sampling of Flexible Ligand-binding Protein Loops

  • Lee, Gyu-Rie;Shin, Woong-Hee;Park, Hahn-Beom;Shin, Seok-Min;Seok, Cha-Ok
    • Bulletin of the Korean Chemical Society
    • /
    • v.33 no.3
    • /
    • pp.770-774
    • /
    • 2012
  • Protein loops are often involved in diverse biological functions, and some functional loops show conformational changes upon ligand binding. Since this conformational change is directly related to ligand binding pose and protein function, there have been numerous attempts to predict this change accurately. In this study, we show that it is plausible to obtain meaningful ensembles of loop conformations for flexible, ligand-binding protein loops efficiently by applying a loop modeling method. The loop modeling method employs triaxial loop closure algorithm for trial conformation generation and conformational space annealing for global energy optimization. When loop modeling was performed on the framework of ligand-free structure, loop structures within $3\AA$ RMSD from the crystal loop structure for the ligand-bound state were sampled in 4 out of 6 cases. This result is encouraging considering that no information on the ligand-bound state was used during the loop modeling process. We therefore expect that the present loop modeling method will be useful for future developments of flexible protein-ligand docking methods.

The Evaluation of Water Vapor Transport and Waterproofness Properties of the Waterproof and Breathable Fabrics (국내 시판 의류용 투습방수소재의 투습방수 성능평가)

  • Shim, Huen Sup
    • The Korean Journal of Community Living Science
    • /
    • v.27 no.2
    • /
    • pp.295-304
    • /
    • 2016
  • This study was conducted to evaluate and compare the waterproofness and water vapor transport properties of 30 waterproof and breathable textiles in the market. MVTR upright cup test with water, MVTR inverted cup test with potassium acetate, and sweating hot plate test method were used for measuring the breathability of fabric samples. The waterproofness of all fabric samples and evaporative resistance of garment ensembles made with 5 selected waterproof and breathable textiles out of 30 samples were tested. The results of this study were as follows. The base fabrics of the most samples were basic weave structure with either polyester or nylon fiber. The waterproofness was over $5000mmH_2O$ for all the samples except densely woven fabrics. The evaporative resistance of garments that were made out of 5 fabric samples with a range of low to high breathability ranged $39.9{\sim}56.7m^2{\cdot}P{\cdot}W^{-1}$. The data of the water vapor transport properties of fabric samples varied with different test methods. Overall the bicomponent film fabrics showed better breathability and less standard deviation regardless of test methods. It is expected the breathability data without the information of test conditions used for marketing would confuse the consumers and the apparel manufacturers.

Seismic response modification factors for stiffness degrading soil-structure systems

  • Ganjavi, Behnoud;Bararnia, Majid;Hajirasouliha, Iman
    • Structural Engineering and Mechanics
    • /
    • v.68 no.2
    • /
    • pp.159-170
    • /
    • 2018
  • This paper aims to develop response modification factors for stiffness degrading structures by incorporating soil-structure interaction effects. A comprehensive parametric study is conducted to investigate the effects of key SSI parameters, natural period of vibration, ductility demand and hysteretic behavior on the response modification factor of soil-structure systems. The nonlinear dynamic response of 6300 soil-structure systems are studied under two ensembles of accelograms including 20 recorded and 7 synthetic ground motions. It is concluded that neglecting the stiffness degradation of structures can results in up to 22% underestimation of inelastic strength demands in soil-structure systems, leading to an unexpected high level of ductility demand in the structures located on soft soil. Nonlinear regression analyses are then performed to derive a simplified expression for estimating ductility-dependent response modification factors for stiffness degrading soil-structure systems. The adequacy of the proposed expression is investigated through sensitivity analyses on nonlinear soil-structure systems under seven synthetic spectrum compatible earthquake ground motions. A good agreement is observed between the results of the predicted and the target ductility demands, demonstrating the adequacy of the expression proposed in this study to estimate the inelastic demands of SSI systems with stiffness degrading structures. It is observed that the maximum differences between the target and average target ductility demands was 15%, which is considered acceptable for practical design purposes.

Classification of Imbalanced Data Using Multilayer Perceptrons (다층퍼셉트론에 의한 불균현 데이터의 학습 방법)

  • Oh, Sang-Hoon
    • The Journal of the Korea Contents Association
    • /
    • v.9 no.7
    • /
    • pp.141-148
    • /
    • 2009
  • Recently there have been many research efforts focused on imbalanced data classification problems, since they are pervasive but hard to be solved. Approaches to the imbalanced data problems can be categorized into data level approach using re-sampling, algorithmic level one using cost functions, and ensembles of basic classifiers for performance improvement. As an algorithmic level approach, this paper proposes to use multilayer perceptrons with higher-order error functions. The error functions intensify the training of minority class patterns and weaken the training of majority class patterns. Mammography and thyroid data-sets are used to verify the superiority of the proposed method over the other methods such as mean-squared error, two-phase, and threshold moving methods.

In vivo Evaluation of Flow Estimation Methods for 3D Color Doppler Imaging

  • Yoo, Yang-Mo
    • Journal of Biomedical Engineering Research
    • /
    • v.31 no.3
    • /
    • pp.177-186
    • /
    • 2010
  • In 3D ultrasound color Doppler imaging (CDI), 8-16 pulse transmissions (ensembles) per each scanline are used for effective clutter rejection and flow estimation, but it yields a low volume acquisition rate. In this paper, we have evaluated three flow estimation methods: autoregression (AR), eigendecomposition (ED), and autocorrelation combined with adaptive clutter rejection (AC-ACR) for a small ensemble size (E=4). The performance of AR, ED and AC-ACR methods was compared using 2D and 3D in vivo data acquired under different clutter conditions (common carotid artery, kidney and liver). To evaluate the effectiveness of three methods, receiver operating characteristic (ROC) curves were generated. For 2D kidney in vivo data, the AC-ACR method outperforms the AR and ED methods in terms of the area under the ROC curve (AUC) (0.852 vs. 0.793 and 0.813, respectively). Similarly, the AC-ACR method shows higher AUC values for 2D liver in vivo data compared to the AR and ED methods (0.855 vs. 0.807 and 0.823, respectively). For the common carotid artery data, the AR provides higher AUC values, but it suffers from biased estimates. For 3D in vivo data acquired from a kidney transplant patient, the AC-ACR with E=4 provides an AUC value of 0.799. These in vivo experiment results indicate that the AC-ACR method can provide more robust flow estimates compared to the AR and ED methods with a small ensemble size.

Characteristics of Export Articles in Korean Clothing Trade -Focused on the 1990's- (한국 수출의류제품의 품목 특성 -1990년대를 중심으로-)

  • Ji, Bye-Kyung
    • Journal of the Korea Fashion and Costume Design Association
    • /
    • v.9 no.1
    • /
    • pp.23-33
    • /
    • 2007
  • Clothing exports of Korea has grown rapidly till the latter half of the 1980's, contributing Korean economic development. However from the 1990's, the amount, the world market share and the international competitiveness of clothing exports have declined. Based on these phenomena, the purpose of this study was to identify the characteristics of export articles in Korean Clothing Trade focused on the 1990's. Statistical data of clothing articles(SITC 84 : Articles of apparel & clothing accessories) were used. The relative importance, trade orientation tendency and unit price of each export clothing articles were analyzed. The results of the study were as follows. On the relative importance, trade orientation tendency and unit price of each export clothing articles, outer garments or products that required complicated production process(e.g., coats, suits, ensembles, jackets, dress) had been decreased in the portion and weakened in the export orientation tendency. But one item in a set or casual wear like trousers, skirts, blouses, shirts, Jerseys, pullovers, T-shirts has been increased in the portion and risen in the unit price. These trends means that clothing exports of Korea were more focused on those category and the international competitiveness on those articles were advanced. From these results, this study can be contributed to establish the concrete clothing export articles strategies of Korean firms.

  • PDF

Differential Effects of Cotton and Polyester Ensembles on Changes in Clothing Surface Temperature, Skin Temperature and Skin Blood Flow During Heat Load

  • Tanaka, Kaori;Hirata, Kozo
    • Fashion & Textile Research Journal
    • /
    • v.2 no.5
    • /
    • pp.405-410
    • /
    • 2000
  • The effects of individual sweating rates on thermophysiological responses during the decrease phase of clothing surface temperature (Tcs) with cotton (C) and polyester (P) clothed subjects were examined. Seven women subjects were exposed in a climatic chamber at ambient temperature of $27.2^{\circ}C$, relative humidity of 50%, and their lower-legs were immersed in a water bath at a temperature of $35-41^{\circ}C$ for 70 min. During water immersion, Tcs in C-clothed subject rose immediately after the onset of sweating and then Tcs fell gradually. In C-clothed subjects, Tcs decreased directly (p<0.05) in proportion to total sweating rate (TSR), however no significant correlation was observed in P-clothed subjects. The relationship between TSR and changes in mean skin temperature, and skin blood flow showed negative correlation when wearing C-clothing (p<0.05), however, no significant correlation when wearing P-clothing. Individual TSR was correlated with threshold rectal temperature for sweating onset (p<0.05) with C- and P-clothed subjects. The results showed that individual TSR had significant effects on not only Tcs but also on thermoregulatory responses during the Tcs decrease phase.

  • PDF