ESTIMATING THE ACTIVATION FUNCTIONS OF AN MLP-NETWORK

P. V. Vehviläinen, H. A. T. Ihalainen

Abstract:

The most common neural network type is a multilayer perceptron (MLP) network. Even a relatively simple MLP-network requires usually high computational accuracy and efficiency. The usual types of activation functions are not necessarily applicable for a neural network algorithm used in microcontrollers. For this purpose a new type of activation function is introduced.
Keywords:

neural network, activation function, estimation, multilayer perceptron
Download:

IMEKO-WC-2000-AI-P595.pdf
DOI:

_unreg_wc-2000.595

Event details:

IMEKO TC:
Event name:

XVI IMEKO World Congress
Title:

Measurement - Supports Science - Improves Technology - Protects Environment ... and Provides Employment - Now and in the Future
Place:

Vienna, AUSTRIA
Time:

25 September 2000 - 28 September 2000