Native Language Recognition using Bidirectional Long Short-Term Memory Network

Speech Recognition of native language is the process of recognizing the language of a client dependent on the speech or content writing in another language. This article proposes the utilization of spectrogram as well as on cochleagram-oriented concepts separated from extremely short speech expressions (0.8 s by and large) to deduce the local language of the speaking person. The bidirectional long short-term memory (BLSTM) neural systems are received to classify the expressions between the local dialects. A lot of analyses is completed for the system engineering look and the framework's precision is assessed on the approval informational index. By and large precision is accomplished utilizing the Mel-recurrence Cepstral coefficients (MRCC) and Gammatone Recurrence Cepstral Coefficients (GRCC), separately. In addition, the advanced MFCC oriented BLSTM system and GFCC based BLSTM systems are combined to make use of their features. The examinations demonstrate that the execution of the combined system outperforms the individual BLSTM systems and precision of 75.69% is accomplished on the assessment information.

Download Full-text

Investigation of Senone-based Long-Short Term Memory RNNs for Spoken Language Recognition

10.21437/odyssey.2016-13 ◽

2016 ◽

Cited By ~ 7

Author(s):

Yao Tian ◽

Liang He ◽

Yi Liu ◽

Jia Liu

Keyword(s):

Short Term Memory ◽

Spoken Language ◽

Language Recognition ◽

Short Term ◽

Term Memory ◽

Long Short Term Memory

Download Full-text

Native Language Identification in Very Short Utterances Using Bidirectional Long Short-Term Memory Network

IEEE Access ◽

10.1109/access.2019.2896453 ◽

2019 ◽

Vol 7 ◽

pp. 17098-17110 ◽

Cited By ~ 2

Author(s):

Farah Adeeba ◽

Sarmad Hussain

Keyword(s):

Short Term Memory ◽

Native Language ◽

Language Identification ◽

Short Term ◽

Term Memory ◽

Memory Network ◽

Long Short Term Memory ◽

Short Utterances

Download Full-text

Short-Term Solar Power Forecasting: A Combined Long Short-Term Memory and Gaussian Process Regression Method

Sustainability ◽

10.3390/su13073665 ◽

2021 ◽

Vol 13 (7) ◽

pp. 3665

Author(s):

Ying Wang ◽

Bo Feng ◽

Qing-Song Hua ◽

Li Sun

Keyword(s):

Gaussian Process ◽

Hybrid Model ◽

Solar Power ◽

Short Term Memory ◽

Gaussian Process Regression ◽

Short Term ◽

Term Memory ◽

Long Short Term Memory ◽

The Individual ◽

Solar Power Forecasting

Solar power is considered a promising power generation candidate in dealing with climate change. Because of the strong randomness, volatility, and intermittence, its safe integration into the smart grid requires accurate short-term forecasting with the required accuracy. The use of solar power should meet requirements proscribed by environmental law and safety standards applied for consumer protection. First, time-series-based solar power forecasting (SPF) model is developed with the time element and predicted weather information from the local meteorological station. Considering the data correlation, long short-term memory (LSTM) algorithm is utilized for short-term SPF. However, the point prediction provided by LSTM fails in revealing the underlying uncertainty range of the solar power output, which is generally needed in some stochastic optimization frameworks. A novel hybrid strategy combining LSTM and Gaussian process regression (GPR), namely LSTM-GPR, is proposed to obtain a highly accurate point prediction with a reliable interval estimation. The hybrid model is evaluated in comparison with other algorithms in terms of two aspects: Point prediction accuracy and interval forecasting reliability. Numerical investigations confirm the superiority of LSTM algorithm over the conventional neural networks. Furthermore, the performance of the proposed hybrid model is demonstrated to be slightly better than the individual LSTM model and significantly superior to the individual GPR model in both point prediction and interval forecasting, indicating a promising prospect for future SPF applications.

Download Full-text

Constant Q Cepstral Coefficients and Long Short-Term Memory Model-Based Automatic Speaker Verification System

Advances in Intelligent Systems and Computing - Proceedings of International Conference on Intelligent Computing, Information and Control Systems ◽

10.1007/978-981-15-8443-5_76 ◽

2021 ◽

pp. 895-904 ◽

Cited By ~ 1

Author(s):

Aakshi Mittal ◽

Mohit Dua

Keyword(s):

Short Term Memory ◽

Speaker Verification ◽

Memory Model ◽

Short Term ◽

Term Memory ◽

Model Based ◽

Verification System ◽

Long Short Term Memory ◽

Cepstral Coefficients

Download Full-text

Sign language recognition with long short-term memory

2016 IEEE International Conference on Image Processing (ICIP) ◽

10.1109/icip.2016.7532884 ◽

2016 ◽

Cited By ~ 14

Author(s):

Tao Liu ◽

Wengang Zhou ◽

Houqiang Li

Keyword(s):

Sign Language ◽

Short Term Memory ◽

Language Recognition ◽

Sign Language Recognition ◽

Short Term ◽

Term Memory ◽

Long Short Term Memory

Download Full-text

Image Caption Generator using CNN-LSTM Deep Neural Network

International Journal for Research in Applied Science and Engineering Technology ◽

10.22214/ijraset.2021.35663 ◽

2021 ◽

Vol 9 (VI) ◽

pp. 2968-2974

Author(s):

Vaibhav Julakanti

Keyword(s):

Neural Network ◽

Deep Neural Network ◽

Short Term Memory ◽

The Other ◽

Short Term ◽

Term Memory ◽

Long Short Term Memory ◽

The Individual ◽

General Climate ◽

Image Caption

Captioning pictures naturally is one of the significant aspects of the human visual framework. There are numerous benefits if there is a model which consequently inscription the scenes or climate encompassed by them and offers back the subtitle as a plain book. In this paper, we present a model dependent on CNN-LSTM neural organizations which naturally identifies the items in the pictures and creates inscriptions for the pictures. It utilizes Inception v3 pre-prepared model to play out the errand of distinguishing items and utilizations LSTM to produce the subtitles. It utilizes the method of Transfer Learning on pre-prepared models for the undertaking of item Detection. This model can perform two activities. The first is to recognize objects in the picture utilizing Convolutional Neural Networks and the other is to subtitle the pictures utilizing RNN based LSTM (Long Short Term Memory). It additionally utilizes a bar look for anticipating the inscriptions for example choosing the best words from the accessible corps. In this, we take top k expectations, feed them again in the model and afterward sort them utilizing the probabilities returned by the model. A portion of the product prerequisites of this undertaking is Tensor Flow V2.0, pandas, NumPy, pickle, PIL, OpenCV. A little GUI is made to transfer the picture to the model to create the inscription. The fundamental use instance of this undertaking is to help outwardly debilitated to comprehend the general climate and act as per that. The inscription age is one of the intriguing and centred fields of Artificial Intelligence which has numerous difficulties to survive. Inscription age includes different complex situations beginning from picking the dataset, preparing the model, approving the model, making pre-prepared models to test the pictures, identifying the pictures lastly producing the individual picture-based subtitles.

Download Full-text