scholarly journals A Self-Supervised Tibetan-Chinese Vocabulary Alignment Method based on Adversarial Learning

2021 ◽  
Author(s):  
Enshuai Hou ◽  
Jie zhu
Keyword(s):  
The Self ◽  
Experimental Results ◽  
Chinese Characters ◽  
Training Method ◽  
Alignment Method ◽  
Adversarial Learning ◽  
Semantic Correlation ◽  
Data Alignment ◽  
Similarity Calculation ◽  
Adversarial Training

Tibetan is a low-resource language. In order to alleviate the shortage of parallel corpus between Tibetan and Chinese, this paper uses two monolingual corpora and a small number of seed dictionaries to learn the semi-supervised method with seed dictionaries and self-supervised adversarial training method through the similarity calculation of word clusters in different embedded spaces and puts forward an improved self-supervised adversarial learning method of Tibetan and Chinese monolingual data alignment only. The experimental results are as follows. First, the experimental results of Tibetan syllables Chinese characters are not good, which reflects the weak semantic correlation between Tibetan syllables and Chinese characters; second, the seed dictionary of semi-supervised method made before 10 predicted word accuracy of 66.5 (Tibetan - Chinese) and 74.8 (Chinese - Tibetan) results, to improve the self-supervision methods in both language directions have reached 53.5 accuracy.

scholarly journals A Self-Supervised Tibetan-Chinese Vocabulary Alignment Method

2021 ◽  
Vol 12 (4) ◽  
pp. 13-20
Author(s):  
Enshuai Hou ◽  
Jie zhu ◽  
Liangcheng Yin ◽  
Ma Ni
Keyword(s):  
The Self ◽  
Experimental Results ◽  
Learning Method ◽  
Training Method ◽  
Alignment Method ◽  
Adversarial Learning ◽  
Parallel Corpus ◽  
Data Alignment ◽  
Similarity Calculation ◽  
Adversarial Training

Tibetan is a low-resource language. In order to alleviate the shortage of parallel corpus between Tibetan and Chinese, this paper uses two monolingual corpora and a small number of seed dictionaries to learn the semi-supervised method with seed dictionaries and self-supervised adversarial training method through the similarity calculation of word clusters in different embedded spaces and puts forward an improved selfsupervised adversarial learning method of Tibetan and Chinese monolingual data alignment only. The experimental results are as follows. The seed dictionary of semi-supervised method made before 10 predicted word accuracy of 66.5 (Tibetan - Chinese) and 74.8 (Chinese - Tibetan) results, to improve the self-supervision methods in both language directions have reached 53.5 accuracy.

scholarly journals Harnessing the Adversarial Perturbation to Enhance Security in the Autoencoder-Based Communication System

Electronics ◽  
2020 ◽  
Vol 9 (2) ◽  
pp. 294
Author(s):  
Zhixiang Deng ◽  
Qian Sang
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Communication System ◽  
Deep Neural Network ◽  
The Self ◽  
Full Duplex ◽  
Training Method ◽  
Adversarial Training ◽  
Simulation Results ◽  
Secure Transmission

Given the vulnerability of deep neural network to adversarial attacks, the application of deep learning in the wireless physical layer arouses comprehensive security concerns. In this paper, we consider an autoencoder-based communication system with a full-duplex (FD) legitimate receiver and an external eavesdropper. It is assumed that the system is trained from end-to-end based on the concepts of autoencoder. The FD legitimate receiver transmits a well-designed adversary perturbation signal to jam the eavesdropper while receiving information simultaneously. To defend the self-perturbation from the loop-back channel, the legitimate receiver is re-trained with the adversarial training method. The simulation results show that with the scheme proposed in this paper, the block-error-rate (BLER) of the legitimate receiver almost remains unaffected while the BLER of the eavesdropper is increased by orders of magnitude. This ensures reliable and secure transmission between the transmitter and the legitimate receiver.

Lateral Tire Forces on Wet Roads

1977 ◽  
Vol 5 (2) ◽  
pp. 75-82 ◽  
Author(s):  
A. Schallamach
Keyword(s):  
Normal Load ◽  
The Self ◽  
Experimental Results ◽  
Slip Angle ◽  
Tire Model ◽  
Side Force ◽  
Tire Forces

Abstract Expressions are derived for side force and self-aligning torque of a simple tire model on wet roads with velocity-dependent friction. The results agree qualitatively with experimental results at moderate speeds. In particular, the theory correctly predicts that the self-aligning torque can become negative under easily realizable circumstances. The slip angle at which the torque reverses sign should increase with the normal load.

scholarly journals Diversity Adversarial Training against Adversarial Attack on Deep Neural Networks

Symmetry ◽  
2021 ◽  
Vol 13 (3) ◽  
pp. 428
Author(s):  
Hyun Kwon ◽  
Jun Lee
Keyword(s):  
Neural Networks ◽  
Deep Neural Networks ◽  
Diversity Training ◽  
Original Data ◽  
Training Method ◽  
Learning Framework ◽  
Adversarial Examples ◽  
Adversarial Training ◽  
Adversarial Attack ◽  
Accuracy Rates

This paper presents research focusing on visualization and pattern recognition based on computer science. Although deep neural networks demonstrate satisfactory performance regarding image and voice recognition, as well as pattern analysis and intrusion detection, they exhibit inferior performance towards adversarial examples. Noise introduction, to some degree, to the original data could lead adversarial examples to be misclassified by deep neural networks, even though they can still be deemed as normal by humans. In this paper, a robust diversity adversarial training method against adversarial attacks was demonstrated. In this approach, the target model is more robust to unknown adversarial examples, as it trains various adversarial samples. During the experiment, Tensorflow was employed as our deep learning framework, while MNIST and Fashion-MNIST were used as experimental datasets. Results revealed that the diversity training method has lowered the attack success rate by an average of 27.2 and 24.3% for various adversarial examples, while maintaining the 98.7 and 91.5% accuracy rates regarding the original data of MNIST and Fashion-MNIST.

SEGMENTATION OF HANDWRITTEN CHINESE CHARACTERS FROM DESTINATION ADDRESSES OF MAIL PIECES

2002 ◽  
Vol 16 (01) ◽  
pp. 85-96 ◽  
Author(s):  
YUE LU ◽  
CHEW LIM TAN ◽  
PENGFEI SHI ◽  
KEHUA ZHANG
Keyword(s):  
Hough Transform ◽  
Structural Features ◽  
Experimental Results ◽  
Chinese Characters ◽  
Topological Relations ◽  
Destination Address

In this paper, we illustrate a method to segment handwritten Chinese characters from destination addresses of mail pieces. Fast Hough transform is utilized to detect the reference lines preprinted on the mail piece. In the segmentation, subassemblies of Chinese characters are merged based on the structural features of Chinese characters and the subassemblies' topological relations, viz. upper–lower, inside–outside and left–right relations. The width of subassemblies and the spacing between neighboring subassemblies in the whole image of the destination address are analyzed to guide the merging of the left–right subassemblies. Experimental results with real mail piece images show that the proposed approach has achieved a promising performance for segmenting handwritten Chinese characters.

RAMAN SPECTRA OF THE H2SO4–SO3 AND D2SO4–SO3 SYSTEMS: EVIDENCE FOR THE OCCURRENCE OF POLYSULPHURIC ACIDS

1962 ◽  
Vol 40 (4) ◽  
pp. 658-674 ◽  
Author(s):  
R. J. Gillespie ◽  
E. A. Robinson
Keyword(s):  
Sulphuric Acid ◽  
Raman Spectra ◽  
Ionic Species ◽  
The Self ◽  
Experimental Results ◽  
Similar Work ◽  
Stretching Vibrations ◽  
Sulphur Trioxide ◽  
High Concentrations

The Raman spectra of oleums, i.e. mixtures of sulphur trioxide and sulphuric acid, have been re-examined. Similar measurements on the sulphur trioxide – deuterosulphuric acid (D2SO4) system are also reported. The experimental results and conclusions of previous similar work on oleums are discussed. By comparison of the spectra of oleums with those of the polysulphuryl halides it is shown that the polysulphuric acids H2S2O7 and H2S3O10 are present in this system. The increase in the frequency of the SO2 stretching vibrations with increasing concentration of sulphur trioxide gives evidence for the existence of higher polysulphuric acids such as H2S4O13 at high concentrations of sulphur trioxide. In relatively concentrated oleum, sulphur trioxide monomer and trimer are also present. It is shown that the self-dissociation of liquid H2S2O7 gives mainly molecular H2S2O10 and H2SO4 and not ionic species. The conclusions reached from the interpretation of the Raman spectra of the D2SO4–SO3 system are similar to those arrived at for sulphuric acid oleums. The spectra of solutions of NaHSO4 in oleums were also examined, and are discussed.

An Analysis Method of Semantic Relevancy on Controlled Tag

2012 ◽  
Vol 263-266 ◽  
pp. 1588-1592
Author(s):  
Jiu Qing Li ◽  
Chi Zhang ◽  
Peng Zhou Zhang
Keyword(s):  
Semantic Similarity ◽  
Calculation Method ◽  
Domain Ontology ◽  
Analysis Method ◽  
Semantic Correlation ◽  
Retrieval Efficiency ◽  
Similarity Calculation

To solve resource-tagging inefficiency and low-precision retrieval in special field, an analysis method of tag semantic relevancy based on controlled database was proposed. The characteristic of special field and building method for controlled database were discussed. Domain ontology correlation calculation method was used to get semantic correlation. The tag semantic similarity calculation method was developed for semantic similarity, and normalization was used to increase the similarity accuracy. With semantic correlation and similarity as parameters, the semantic relevancy in special field can be obtained. This method was used successfully in the special field of actual projects, improved resource-tagging and retrieval efficiency.

scholarly journals Accelerating Computation of Zernike and Pseudo-Zernike Moments with a GPU Algorithm

2020 ◽  
Author(s):  
◽  
Shiying Li
Keyword(s):  
Character Recognition ◽  
Signal To Noise Ratio ◽  
Zernike Moments ◽  
Experimental Results ◽  
Chinese Characters ◽  
Maximum Order ◽  
Practical Applications ◽  
Feature Vectors ◽  
Computation Process ◽  
Image Reconstructions

Although Zernike and pseudo-Zernike moments have some advanced properties, the computation process is generally very time-consuming, which has limited their practical applications. To improve the computational efficiency of Zernike and pseudo-Zernike moments, in this research, we have explored the use of GPU to accelerate moments computation, and proposed a GPUaccelerated algorithm. The newly developed algorithm is implemented in Python and CUDA C++ with optimizations based on symmetric properties and k × k sub-region scheme. The experimental results are encouraging and have shown that our GPU-accelerated algorithm is able to compute Zernike moments up to order 700 for an image sized at 512 × 512 in 1.7 seconds and compute pseudo-Zernike moments in 3.1 seconds. We have also verified the accuracy of our GPU algorithm by performing image reconstructions from the higher orders of Zernike and pseudo-Zernike moments. For an image sized at 512 × 512, with the maximum order of 700 and k = 11, the PSNR (Peak Signal to Noise Ratio) values of its reconstructed versions from Zernike and pseudo-Zernike moments are 44.52 and 46.29 separately. We have performed image reconstructions from partial sets of Zernike and pseudo-Zernike moments with various order n and different repetition m. Experimental results of both Zernike and pseudo-Zernike moments show that the images reconstructed from the moments of lower and higher orders preserve the principle contents and details of the original image respectively, while moments of positive and negative m result in identical images. Lastly, we have proposed a set of feature vectors based on pseudo-Zernike moments for Chinese character recognition. Three different feature vectors are composed of different parts of four selected lower pseudo-Zernike moments. Experiments on a set of 6,762 Chinese characters show that this method performs well to recognize similar-shaped Chinese characters.

scholarly journals Experimental Study and Extended Finite Element Simulation of Fracture of Self-Compacting Rubberized Concrete

2021 ◽  
Vol 2021 ◽  
pp. 1-12
Author(s):  
Xinquan Wang ◽  
Hongguo Diao ◽  
Yunliang Cui ◽  
Changguang Qi ◽  
Shangyu Han
Keyword(s):  
Finite Element ◽  
Loading Rate ◽  
High Performance Concrete ◽  
The Self ◽  
Experimental Results ◽  
Crack Mouth Opening Displacement ◽  
Displacement Curve ◽  
Rubberized Concrete ◽  
Extended Finite Element ◽  
Self Compacting Concrete

Self-compacting rubberized concrete (SCRC) is a high-performance concrete that can achieve compacting effect by self-gravity without vibration during pouring. Because of its excellent fluidity, homogeneity, and stability, the application of self-compacting concrete in engineering can improve work efficiency and reduce project cost. The effects of loading rate on the fracture behavior of self-compacting concrete were studied in this paper. Three-point bend (TPB) tests were carried out at five loading rates of 1, 0.1, 0.001, 0.0001, and 0.00001 mm/s. The dimensions of the specimens were 100  mm × 100 mm × 400 mm. A precast crack was set in the middle of the specimen with a notch-depth ratio of 0.4. The experimental results show that the peak load on the load-CMOD (crack mouth opening displacement) curve gradually increases with the increase of the loading rate. Although the fracture energy a presented greater dispersion under the loading rate of 1 mm/s, the overall changes were still rising with the increase of the loading rate. Besides studying the softening characteristics of the self-compacting concrete, the constitutive softening curve of the self-compacting concrete was obtained using the bilinear model. Finally, curved three-point bending beams were simulated by using the extended finite element method based on ABAQUS. The fracture process of the self-compacting concrete under different loading conditions was analyzed more intuitively. The simulation results were compared with the experimental results, and the same conclusions were obtained.

Sign in / Sign up

Export Citation Format

Share Document