Method for EMR and AE interference signal identification in coal rock mining based on recurrent neural networks

2021 ◽  
Author(s):  
Yangyang Di ◽  
Enyuan Wang ◽  
Zhonghui Li ◽  
Xiaofei Liu ◽  
Baolin Li
Keyword(s):  
Neural Networks ◽  
Recurrent Neural Networks ◽  
Interference Signal ◽  
Signal Identification ◽  
Rock Mining ◽  
Coal Rock

EMR Interference Signal Recognition in Coal Rock Mining Based on Recurrent Neural Networks

Geophysics ◽  
2021 ◽  
pp. 1-59
Author(s):  
yangyang Di ◽  
Enyuan Wang
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Rock Burst ◽  
Recurrent Neural Networks ◽  
Short Term Memory ◽  
Underground Mining ◽  
Interference Signal ◽  
Signal Recognition ◽  
Interference Signals ◽  
Coal Rock

The electromagnetic radiation (EMR) method is a promising geophysical method for monitoring and providing early warnings about coal rock burst disasters. In the underground mining process, personnel activities and electromechanical equipment produce EMR interference signals that affect the accuracy of EMR monitoring. Current methods for identifying the EMR interference signals mainly use their time and amplitude characteristics. However, these methods of EMR interference signal recognition and filtering need to be further improved. The advancements in the deep learning technique provide an opportunity to develop a new method for their identification and filtering. A method for EMR interference signal recognition based on deep learning algorithms is proposed. The proposed method uses bidirectional long short-term memory recurrent neural networks and Fourier transform to analyze numerous EMR interference signals along with other signals to intelligently identify and filter EMR signal sequences. The results showed that the proposed method can respond positively to EMR interferences and accurately eliminate EMR interference signals. This method can significantly improve the reliability of EMR monitoring data and effectively monitor rock burst disasters.

Levenshtein Augmentation Improves Performance of SMILES Based Deep-Learning Synthesis Prediction

2020 ◽  
Author(s):  
Dean Sumner ◽  
Jiazhen He ◽  
Amol Thakkar ◽  
Ola Engkvist ◽  
Esben Jannik Bjerrum
Keyword(s):  
Neural Networks ◽  
Pattern Recognition ◽  
Deep Learning ◽  
Recurrent Neural Networks ◽  
Data Augmentation ◽  
State Of The Art ◽  
Sequence Similarity ◽  
Learning Models ◽  
Underlying Network

<p>SMILES randomization, a form of data augmentation, has previously been shown to increase the performance of deep learning models compared to non-augmented baselines. Here, we propose a novel data augmentation method we call “Levenshtein augmentation” which considers local SMILES sub-sequence similarity between reactants and their respective products when creating training pairs. The performance of Levenshtein augmentation was tested using two state of the art models - transformer and sequence-to-sequence based recurrent neural networks with attention. Levenshtein augmentation demonstrated an increase performance over non-augmented, and conventionally SMILES randomization augmented data when used for training of baseline models. Furthermore, Levenshtein augmentation seemingly results in what we define as <i>attentional gain </i>– an enhancement in the pattern recognition capabilities of the underlying network to molecular motifs.</p>

LatticeRnn: Recurrent Neural Networks Over Lattices

2016 ◽  
Author(s):  
Faisal Ladhak ◽  
Ankur Gandhe ◽  
Markus Dreyer ◽  
Lambert Mathias ◽  
Ariya Rastrow ◽  
...  
Keyword(s):  
Neural Networks ◽  
Recurrent Neural Networks
Sign in / Sign up

Export Citation Format

Share Document