SSIT: a sample selection-based incremental model training method for image recognition

2021 ◽  
Author(s):  
Yichuan Zhang ◽  
Yadi Liu ◽  
Guangming Yang ◽  
Jie Song
Keyword(s):  
Image Recognition ◽  
Sample Selection ◽  
Training Method ◽  
Incremental Model ◽  
Model Training

Federated Learning Model Training Method Based on Data Features Perception Aggregation

2021 ◽  
Author(s):  
Zeng Yan ◽  
Yan Zhong Yi ◽  
Zhang JiLin ◽  
Zhao NaiLiang ◽  
Ren YongJian ◽  
...  
Keyword(s):  
Learning Model ◽  
Training Method ◽  
Model Training

Improving the Performance of Language Identification System Using Different GMM-Training Approaches

2010 ◽  
Vol 121-122 ◽  
pp. 496-501
Author(s):  
Wei Li ◽  
Dong Ju Kim ◽  
Kwang Seok Hong
Keyword(s):  
System Performance ◽  
Linear Prediction ◽  
Gaussian Mixture ◽  
High Accuracy ◽  
Language Identification ◽  
Identification System ◽  
Training Method ◽  
Model Training ◽  
Optimal Approach ◽  
Feasible System

This paper proposed a feasible system for language identification (LID) and designed four different GMM-training approaches to improve the system performance by accuracy recognition rates. In our experiment, we used these model-training approaches to evaluation on the system performance, which utilizes Linear Prediction Cepstrum Coefficients (LPCC) and Gaussian Mixture Model (GMM), rely on a 10-language task. From all the results, we found an optimal approach for training GMM in LID system, which achieves high accuracy of 85.25%, and indicated that different GMM-training approaches have different performances for LID system, but an advisable training method that proposed in our paper can greatly improve the system performance.

scholarly journals Pengaruh Latihan Squat Thrust dan Throw Ball Medicine Terhadap Kemampuan Jauhnya Lemparan Chest Pass Bola Basket Ekstrakurikuler SMPN 22 Putri Hijau Bengkulu Utara

2020 ◽  
Vol 1 (2) ◽  
pp. 16-22
Author(s):  
Nuraini Nuraini ◽  
Tono Sugihartono ◽  
Ari Sutisyana
Keyword(s):  
Data Collection ◽  
Sample Selection ◽  
Test Method ◽  
Distributed Data ◽  
Direct Test ◽  
Training Method ◽  
Basketball Court ◽  
Sample Data ◽  
Comparison Groups ◽  
Improve Student Achievement

This study aims to determine the effect of Squat Thrust Training and Throw Ball Medicine on the Ability of the Throwing Basketball Chest Pass. This research was conducted on the basketball court of SMP Negeri 22 Putri Hijau North Bengkulu. in students who take basketball extracurricular. This study uses an experimental method, the design used is the design of one group pretest - posttest that is pretest before given treatment and posttest after being treated. The treatment given there are two forms of exercise namely Squat Thrust and Throw Ball Medicine. This design requires one group without comparison groups. The population in this study amounted to 25 students, sample selection using total sampling where the entire population was taken as a research sample. Data collection techniques in this study with the direct test method that is using the basketball throw test. Statistical prerequisite tests meet homogeneous requirements and normally distributed data based on statistical calculations obtained from the data t count = 8.803> t table = 2.064 with the level ? = 0.05. in the Test of the contribution of Squat Thrust and Throw ball medicine by 18.32% The results of this study indicate that there is an influence between squat thrust training and throw ball medicine on the ability of the extracurricular basketball chest throws of SMP Negeri 22 Putri Hijau North Bengkulu, so it can be concluded that Squat training Thrust and Throw Ball Medicine affect the distance of the chest pass and can be used as a training method to improve student achievement.

scholarly journals A Method for Image Recognition of Insulator Jacket Defects under Small Sample Conditions

2022 ◽  
Vol 2160 (1) ◽  
pp. 012062
Author(s):  
Xinhai Li ◽  
Lingcheng Zeng ◽  
Yongyin Lu ◽  
Yuede Lin ◽  
Xinxiong Zeng
Keyword(s):  
Image Recognition ◽  
Original Data ◽  
Small Sample ◽  
Small Scale ◽  
Learning Technology ◽  
Accurate Identification ◽  
Data Set ◽  
Meta Learning ◽  
Model Training ◽  
Defect Recognition

Abstract Accurate identification of insulator jacket defect images requires a large number of samples for model training, and the actual defect image datasets available for model training is seriously insufficient. In order to solve the problems of the model cannot be trained, over-fitting and low accuracy caused by too few training samples, this paper proposes a new method for image recognition of insulator jacket defects under small sample conditions, which combines image enhancement technology and meta-learning technology to train the U-Net image segmentation network, and finally obtain the image recognition model of the insulator jacket defect. In this paper, the defect recognition models using meta-learning method and without meta-learning are compared experimentally, and the results show that the proposed method can achieve accurate recognition with a small-scale original data set.

scholarly journals An Easy Method for Identifying 315 Categories of Commonly-Used Chinese Herbal Medicines Based on Automated Image Recognition Using AutoML Platforms

2020 ◽  
Author(s):  
Wenting Chen ◽  
Jiayun Tong ◽  
Rui He ◽  
Peiting Chen ◽  
Zixin Chen ◽  
...  
Keyword(s):  
Image Recognition ◽  
Herbal Medicines ◽  
Model Performance ◽  
Easy Method ◽  
Chinese Herbal Medicines ◽  
Training Time ◽  
Chinese Herbal ◽  
Significant Difference ◽  
Model Training ◽  
The One

Abstract Background: The identification and authentication of Chinese herbal medicines (CHMs) are directly related to their safety and efficacy in clinical treatment. However, the limited number of qualified professionals with expertise fails to meet the demand of the vast CHMs market. To make the CHMs identification more convenient and accurate, this study aimed at assessing the feasibility of the state-of-art automated machine learning (AutoML) technology in CHMs image recognition.Methods: This study presented an experimental AutoML model built on the one-stop Huawei ModelArts platform instead of a handcrafted neural network. A rich and representative dataset of 31,460 images consisting of 315 categories of commonly-used CHMs was built and used for the model creation. Furthermore, the Huawei ModelArts model was compared with a model built on the Baidu EasyDL platform using the same dataset to investigate their ability to recognize CHMs images. Three professionals were also invited to recognize images of 315 categories of CHMs.Results: During the model evaluation, high accuracies of 99.2% and 98.4% were achieved by ModelArts and EasyDL, respectively. In the subsequent held-out tests, the accuracies of ModelArts and EasyDL models were 91.2% and 91.85%, respectively. Both models performed very well individually and no statistically significant difference was found in model performance between these two platforms. However, the model-training time was only approximately 41 minutes on ModelArts platform but 118 minutes on EasyDL. The mean accuracy of the manual recognition for 315 CHMs was 97.46±1.58%.Conclusion: Results revealed that AutoML technology is a fast and simple approach and has great practical potential in the field of CHMs image recognition. Since the Huawei ModelArts platform requires less training time, we recommend it as a priority.

scholarly journals STRUCTURE OPTIMIZATION OF DEEP BELIEF NETS IN THE APPLICATIONS OF IMAGE RECOGNITION.

2020 ◽  
Vol 9 (7) ◽  
pp. 179-189
Keyword(s):  
Deep Learning ◽  
Image Recognition ◽  
Structure Optimization ◽  
Reconstruction Error ◽  
Deep Belief Network ◽  
Training Method ◽  
Belief Network ◽  
Supervised Training ◽  
Training Resources ◽  
Hidden Layer

Deep belief network (DBN) has become one of the most important models in deep learning, however, the un-optimized structure leads to wasting too much training resources. To solve this problem and to investigate the connection of depth and accuracy of DBN, an optimization training method that consists of two steps is proposed. Firstly, by using mathematical and biological tools, the significance of supervised training is analyzed, and a theorem, that is on reconstruction error and network energy, is proved. Secondly, based on conclusions of step one, this paper proposes to optimize the structure of DBN (especially hidden layer numbers). Thirdly, this method is applied in two image recognition experiments, and results show increased computing efficiency and accuracies in both tasks.

Sentiment analysis via dually-born-again network and sample selection

2020 ◽  
Vol 24 (6) ◽  
pp. 1257-1271
Author(s):  
Pinlong Zhao ◽  
Zefeng Han ◽  
Qing Yin ◽  
Shuxiao Li ◽  
Ou Wu
Keyword(s):  
Sentiment Analysis ◽  
Language Processing ◽  
State Of The Art ◽  
Sample Selection ◽  
Learning Framework ◽  
Training Samples ◽  
Knowledge Distillation ◽  
Model Training ◽  
Born Again ◽  
Text Sentiment Analysis

Text sentiment analysis is an important natural language processing (NLP) task and has received considerable attention in recent years. Numerous deep-learning based methods have been proposed in previous literature in terms of new deep neural networks (DNN) including new embedding strategies, new attention mechanisms, and new encoding layers. In this study, an alternative technical path is investigated to further improve the state-of-the-art performance of text sentiment analysis. An new effective learning framework is proposed that combines knowledge distillation and sample selection. A dually-born-again network (DBAN) is presented in which the teacher network and the student network are simultaneously trained through an iterative approach. A selection gate is defined to deal with training samples which are useless or even harmful for model training. Moreover, both the DBAN and sample selection are further improved by ensemble. The proposed framework can improve the existing state-of-the-art DNN models in sentiment analysis. Experimental results indicate that the proposed framework enhances the performances of existing networks. In addition, DBAN outperforms existing born-again network.

Sign in / Sign up

Export Citation Format

Share Document