A Semantic Textual Similarity Calculation Model Based on Pre-training Model

2021 ◽  
pp. 3-15
Author(s):  
Zhaoyun Ding ◽  
Kai Liu ◽  
Wenhao Wang ◽  
Bin Liu
Keyword(s):  
Training Model ◽  
Calculation Model ◽  
Model Based ◽  
Similarity Calculation ◽  
Semantic Textual Similarity

scholarly journals IDSSIM: an lncRNA functional similarity calculation model based on an improved disease semantic similarity method

2020 ◽  
Vol 21 (1) ◽  
Author(s):  
Wenwen Fan ◽  
Junliang Shang ◽  
Feng Li ◽  
Yan Sun ◽  
Shasha Yuan ◽  
...  
Keyword(s):  
Semantic Similarity ◽  
Calculation Model ◽  
Functional Similarity ◽  
Model Based ◽  
Similarity Calculation ◽  
Similarity Method

scholarly journals Traditional Chinese Medicine Text Similarity Calculation Model Based on the Bidirectional Temporal Siamese Network

2021 ◽  
Vol 2021 ◽  
pp. 1-10
Author(s):  
Jigen Luo ◽  
Wangping Xiong ◽  
Jianqiang Du ◽  
Yingfeng Liu ◽  
Jianwen Li ◽  
...  
Keyword(s):  
Chinese Medicine ◽  
Traditional Chinese Medicine ◽  
Knowledge Integration ◽  
Short Term Memory ◽  
Calculation Model ◽  
Word Segmentation ◽  
Text Similarity ◽  
Model Based ◽  
Siamese Network ◽  
Similarity Calculation

The text similarity calculation plays a crucial role as the core work of artificial intelligence commercial applications such as traditional Chinese medicine (TCM) auxiliary diagnosis, intelligent question and answer, and prescription recommendation. However, TCM texts have problems such as short sentence expression, inaccurate word segmentation, strong semantic relevance, high feature dimension, and sparseness. This study comprehensively considers the temporal information of sentence context and proposes a TCM text similarity calculation model based on the bidirectional temporal Siamese network (BTSN). We used the enhanced representation through knowledge integration (ERNIE) pretrained language model to train character vectors instead of word vectors and solved the problem of inaccurate word segmentation in TCM. In the Siamese network, the traditional fully connected neural network was replaced by a deep bidirectional long short-term memory (BLSTM) to capture the contextual semantics of the current word information. The improved similarity BLSTM was used to map the sentence that is to be tested into two sets of low-dimensional numerical vectors. Then, we performed similarity calculation training. Experiments on the two datasets of financial and TCM show that the performance of the BTSN model in this study was better than that of other similarity calculation models. When the number of layers of the BLSTM reached 6 layers, the accuracy of the model was the highest. This verifies that the text similarity calculation model proposed in this study has high engineering value.

scholarly journals Evaluation of an intraoperative ultrasound training model based on a cadaveric sheep brain

2014 ◽  
Vol 5 (1) ◽  
pp. 46 ◽  
Author(s):  
Jan Vavruska ◽  
Ralf Buhl ◽  
AthanasiosK Petridis ◽  
Homajoun Maslehaty ◽  
Martin Scholz
Keyword(s):  
Intraoperative Ultrasound ◽  
Training Model ◽  
Model Based ◽  
Sheep Brain

Sports Training Model Based on GA Optimized Neural Network

2020 ◽  
Author(s):  
Xianqiong Ding ◽  
Jin Tang ◽  
Yuwen Zhang ◽  
Zhonghua Yuan
Keyword(s):  
Neural Network ◽  
Training Model ◽  
Sports Training ◽  
Model Based

scholarly journals Research on Tuning Fork Dimension Optimization and Density Calculation Model Based on Viscosity Compensation for Tuning Fork Density Sensor

2020 ◽  
Vol 2020 ◽  
pp. 1-17
Author(s):  
Hai Yang ◽  
Yue Rao ◽  
Li Li ◽  
Haibo Liang ◽  
Tao Luo ◽  
...  
Keyword(s):  
Real Time ◽  
Measurement Accuracy ◽  
Tuning Fork ◽  
High Viscosity ◽  
Calculation Model ◽  
Online Measurement ◽  
Liquid Viscosity ◽  
Density Calculation ◽  
Model Based ◽  
Measurement Resolution

At present, real-time online measurement of fluid density is of great significance to improve the automation level of petrochemical and food industries. The tuning fork density sensor is widely used because of its characteristics of real-time online measurement, high measurement accuracy, simple structure, and convenient use. The traditional tuning fork density sensor in the market has the disadvantage of low resolution and being susceptible to liquid viscosity, which makes the sensor’s measurement accuracy low and not suitable for the measurement of high-viscosity liquid density. The measurement resolution and antiviscosity interference capability of the tuning fork density sensor are two major indexes to measure the measurement performance of the sensor, among the antiviscosity interference capability refers to the degree to which the measurement results of the sensor are affected by viscosity properties. However, the structural design of the tuning fork density sensor results in the conflict between the measurement resolution and the antiviscosity interference capability of the sensor, and the improvement of one performance is bound to affect the performance of the other. Aiming at the problem of how to balance the measuring performance of the tuning fork sensor, a density calculation model based on viscosity compensation is proposed in this paper. By studying the working principle and structure design of the tuning fork, the vibration characteristics of tuning fork in liquid with different viscosities and densities are modelled and simulated. From the results of simulation analysis, the better set of dimensions with balanced measuring performance is selected. Not only does the structure of the tuning fork have the characteristics of high resonance frequency, but also the measured results are less affected by the viscosity of the liquid. To solve the problem that density measurement is still affected by high-viscosity liquid after tuning fork dimension optimization, in this paper, the partial least square model is used to fit the experimental data of the frequency-density characteristics; then, the density calculation model based on the viscosity compensation is obtained by combining the frequency-viscosity characteristic experiment. Finally, through the performance test experiment comparing with the traditional tuning fork density sensor, the measurement resolution of the improved tuning fork density sensor is as high as 0.0001 g/cm3; within the viscosity range of 180 MPa·s, the accuracy reached ±0.001 g/cm3, and within 480 MPa·s, the measurement accuracy reached ±0.002 g/cm3. When the liquid viscosity reaches more than 10 MPa·s, the improved tuning fork density sensor has better overall measurement performance than the traditional tuning fork density sensor, and both of its measurement resolution and antiviscosity interference capability have been greatly improved.

Sign in / Sign up

Export Citation Format

Share Document