Configurable Model for Sigmoid and Hyperbolic Tangent Functions
Recurrent neural networks (RNNs) are considered to be among the most important types of neural networks especially for the applications where processing of a sequence of data comes to place. RNNs are in general computationally expensive and need a lot of processing time and power. Therefore, there i...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
ARQII PUBLICATION
2023-07-01
|
| Series: | Applications of Modelling and Simulation |
| Subjects: | |
| Online Access: | http://arqiipubl.com/ojs/index.php/AMS_Journal/article/view/395/153 |
| _version_ | 1797790007287611392 |
|---|---|
| author | Khaled Salah Mona Safar Mohamed Taher Ashraf Salem |
| author_facet | Khaled Salah Mona Safar Mohamed Taher Ashraf Salem |
| author_sort | Khaled Salah |
| collection | DOAJ |
| description | Recurrent neural networks (RNNs) are considered to be among the most important types of neural networks especially for the applications where processing of a sequence of data comes to place. RNNs are in general computationally expensive and need a lot of processing time and power. Therefore, there is a strong need to reduce the processing time to be able to use them in an embedded environment with limited resources. In this work, we present an accelerated field programmable gate array (FPGA) model for RNNs with an emphasis on long short-term memory neural networks (LSTMs). A new configurable block capable of calculating Tanh and Sigmoid activation functions is proposed and analyzed. The solution is based on a look-up table and additional simple math operations, which leads to a speedup of the proposed model of the neural network. Results are obtained and compared with other work by the simulation tool ISE Xillinx. |
| first_indexed | 2024-03-13T01:58:38Z |
| format | Article |
| id | doaj.art-b2310b78307c4cc0acdd416e57adfdf5 |
| institution | Directory Open Access Journal |
| issn | 2600-8084 |
| language | English |
| last_indexed | 2024-03-13T01:58:38Z |
| publishDate | 2023-07-01 |
| publisher | ARQII PUBLICATION |
| record_format | Article |
| series | Applications of Modelling and Simulation |
| spelling | doaj.art-b2310b78307c4cc0acdd416e57adfdf52023-07-02T08:30:36ZengARQII PUBLICATIONApplications of Modelling and Simulation2600-80842023-07-0177884Configurable Model for Sigmoid and Hyperbolic Tangent FunctionsKhaled Salah0Mona Safar1Mohamed Taher2Ashraf Salem3Computer and Systems Engineering Department, Ain Shams University, Cairo, EgyptComputer and Systems Engineering Department, Ain Shams University, Cairo, EgyptComputer and Systems Engineering Department, Ain Shams University, Cairo, EgyptComputer and Systems Engineering Department, Ain Shams University, Cairo, EgyptRecurrent neural networks (RNNs) are considered to be among the most important types of neural networks especially for the applications where processing of a sequence of data comes to place. RNNs are in general computationally expensive and need a lot of processing time and power. Therefore, there is a strong need to reduce the processing time to be able to use them in an embedded environment with limited resources. In this work, we present an accelerated field programmable gate array (FPGA) model for RNNs with an emphasis on long short-term memory neural networks (LSTMs). A new configurable block capable of calculating Tanh and Sigmoid activation functions is proposed and analyzed. The solution is based on a look-up table and additional simple math operations, which leads to a speedup of the proposed model of the neural network. Results are obtained and compared with other work by the simulation tool ISE Xillinx.http://arqiipubl.com/ojs/index.php/AMS_Journal/article/view/395/153activation functionsdeep learningfpgalstmrecurrent neural network |
| spellingShingle | Khaled Salah Mona Safar Mohamed Taher Ashraf Salem Configurable Model for Sigmoid and Hyperbolic Tangent Functions Applications of Modelling and Simulation activation functions deep learning fpga lstm recurrent neural network |
| title | Configurable Model for Sigmoid and Hyperbolic Tangent Functions |
| title_full | Configurable Model for Sigmoid and Hyperbolic Tangent Functions |
| title_fullStr | Configurable Model for Sigmoid and Hyperbolic Tangent Functions |
| title_full_unstemmed | Configurable Model for Sigmoid and Hyperbolic Tangent Functions |
| title_short | Configurable Model for Sigmoid and Hyperbolic Tangent Functions |
| title_sort | configurable model for sigmoid and hyperbolic tangent functions |
| topic | activation functions deep learning fpga lstm recurrent neural network |
| url | http://arqiipubl.com/ojs/index.php/AMS_Journal/article/view/395/153 |
| work_keys_str_mv | AT khaledsalah configurablemodelforsigmoidandhyperbolictangentfunctions AT monasafar configurablemodelforsigmoidandhyperbolictangentfunctions AT mohamedtaher configurablemodelforsigmoidandhyperbolictangentfunctions AT ashrafsalem configurablemodelforsigmoidandhyperbolictangentfunctions |