• Journal of KIISE
• /
• v.44 no.1
• /
• pp.36-43
• /
• 2017

Syntactic information represents the dependency relation between predicates and arguments, and it is helpful for improving the performance of Semantic Role Labeling systems. However, syntax analysis can cause computational overhead and inherit incorrect syntactic information. To solve this problem, we exclude syntactic information and use only morpheme information to construct Semantic Role Labeling systems. In this study, we propose an end-to-end SRL system that only uses morpheme information with Stacked Bidirectional LSTM-CRFs model by extending the LSTM RNN that is suitable for sequence labeling problem. Our experimental results show that our proposed model has better performance, as compare to other models.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.5
• /
• pp.662-667
• /
• 2022

Semantic role labeling task(SRL) is to extract predicate and arguments such as agent, patient, place, time. In the previously SRL task studies, a pipeline method extracting linguistic features of sentence has been proposed, but in this method, errors of each extraction work in the pipeline affect semantic role labeling performance. Therefore, methods using End-to-End neural network model have recently been proposed. In this paper, we propose a neural network model using the Biaffine Average Attention model for SRL task. The proposed model consists of a structure that can focus on the entire sentence information regardless of the distance between the predicate in the sentence and the arguments, instead of LSTM model that uses the surrounding information for prediction of a specific token proposed in the previous studies. For evaluation, we used F1 scores to compare two models based BERT model that proposed in existing studies using F1 scores, and found that 76.21% performance was higher than comparison models.

• Research in Mathematical Education
• /
• v.19 no.1
• /
• pp.81-88
• /
• 2015

This study used a posttest control group design and to find out differences between students' self-regulated learning strategies in traditional and non-traditional classroom. To this end, 131 first year university students within the experimental and control groups took part in the study. While ICT-based approach was used as the main medium of instruction in the experimental group, in the control group the paper-based traditional method was used. A survey adapted from Davaanyam [Davaanyam, T. (2013). The structural relationships among Mongolian students' attitudes toward mathematics, motivational beliefs, self-regulated learning strategies, and mathematics achievement. Ph. D. Dissertation. Jeonju, Jeonbuk, Korea: Chonbuk National Unversity.] was used to gather the data. The results of the study indicated a significant difference between the control and experimental groups in regard with their self-regulated learning. That is to say, the experimental group taught through ICT tools acquired higher levels of self-regulation as compared with the control group instructed through the traditional teaching method.

• Radiation Oncology Journal
• /
• v.33 no.2
• /
• pp.149-154
• /
• 2015

Purpose: We performed exome sequencing in a breast cancer family without BRCA mutations. Materials and Methods: A family that three sisters have a history of breast cancer was selected for analysis. There were no family members with breast cancer in the previous generation. Genetic testing for BRCA mutation was negative, even by the multiplex ligation-dependent probe amplification method. Two sisters with breast cancer were selected as affected members, while the mother of the sisters was a non-affected member. Whole exome sequencing was performed on the HiSeq 2000 platform with paired-end reads of 101 bp in the three members. Results: We identified 19,436, 19,468, and 19,345 single-nucleotide polymorphisms (SNPs) in the coding regions. Among them, 8,759, 8,789, and 8,772 were non-synonymous SNPs, respectively. After filtering out 12,843 synonymous variations and 12,105 known variations with indels found in the dbSNP135 or 1000 Genomes Project database, we selected 73 variations in the samples from the affected sisters that did not occur in the sample from the unaffected mother. Using the Sorting Intolerant From Tolerant (SIFT), PolyPhen-2, and MutationTaster algorithms to predict amino acid substitutions, the XCR1, DLL1, TH, ACCS, SPPL3, CCNF, and SRL genes were risky among all three algorithms, while definite candidate genes could not be conclusively determined. Conclusion: Using exome sequencing, we found 7 variants for a breast cancer family without BRCA mutations. Genetic evidence of disease association should be confirmed by future studies.

• Asian-Australasian Journal of Animal Sciences
• /
• v.16 no.11
• /
• pp.1610-1613
• /
• 2003

The aim of the experiment was to study the changes in the activities of various rumen fibre degrading enzymes due to the feeding of chemically treated mustard (Brassica campestris) straw in sheep. Mustard straw (MS) (<5 cm particle size) was treated either with urea (4% (w/w), or with 2% sodium hydroxide (NaOH), or with alkaline hydrogen peroxide (2% NaOH and 1.5% hydrogen peroxide ($H_2O_2$)) and/or supplemented with 2% (w/w) urea. Seven maintenance type rations were prepared using MS (70 parts) with molasses (5 parts) and concentrate (25 parts). They were untreated MS (CMS), urea treated MS (UMS), urea supplemented MS (MSUS), alkali treated MS (AMS), alkali treated and urea supplemented MS (AMS-US), alkali $H_2O_2$ treated MS (AHMS) and alkali $H_2O_2$ treated and urea supplemented MS (AHMS-US). They were then compressed into a complete feed block with the help of block making machine. Forty two male hoggets of Malpura breed sheep were equally distributed into each treatment group and (were) offered feed and water ad libitum. At the end of 21 days of feeding trial, rumen liquor was collected through stomach tube from three animals in each group at 0 h, 4 h, 8 h, 12 h of post feeding. Results showed that the level of enzyme varied from 8.52 to 11.12, 40.85 to 50.37, 3.22 to 3.78, 2.09 to 2.77 and 31.44 to 44.24 units/100 ml SRL respectively for carboxymethyl cellulase (CMCase), $\alpha$-amylase, microcrystalline cellulase (MCCase), filter paper (FP) degrading enzyme and $\alpha$-glucosidase. Processing of MS affected the enzyme activities, in a way, that NaOH and AHP treatment significantly reduced CMCase and FP degrading enzyme. The effect of urea treatment showed an increase in the activity of MCCase and $\alpha$-glucosidase. But the supplementation of urea increased the activity of CMCase, FP degrading enzyme and $\alpha$-glucosidase. The CMCase, $\alpha$-amylase, $\alpha$-glucosidase activities were highest at 4hr whereas MCCase and FP degrading enzyme had maximum activities at 12 h post feeding Results suggested that MS might need longer time in the rumen for its effective degradation.