Improving Neural Network Efficiency Using Piecewise Linear Approximation of Activation Functions
Activation functions play a pivotal role in Neural Networks by enabling the modeling of complex non-linear relationships within data. However, the computational cost associated with certain activation functions, such as the hyperbolic tangent (tanh) and its gradient, can be substantial. In this stud...
Saved in:
| Published in: | Proceedings of the International Florida Artificial Intelligence Research Society Conference Vol. 38; no. 1 |
|---|---|
| Main Authors: | , |
| Format: | Journal Article |
| Language: | English |
| Published: |
LibraryPress@UF
14.05.2025
|
| ISSN: | 2334-0754, 2334-0762 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Be the first to leave a comment!