Double-exponential sigmoidal functions for neural networksDoppelexponentielle Sigmoidfunktionen für Neuronale Netze

被引：0

作者：

M. Heiss

S. Kampl

机构：

来源：

e&i Elektrotechnik und Informationstechnik | 1997年 / 114卷 / 7-8期

关键词：

neural network; activation function; sigmoid, no multiplication; double-exponential function; numeric simplicity; Neuronal Netze; Transferfunktion; Sigmoidfunktion; multiplikationsfrei; doppel-exponentiell; Recheneffizienz;

D O I：

10.1007/BF03159051

中图分类号：

学科分类号：

摘要：

A computationally efficient sigmoidal activation function is presented, called a double-exponential signal function, and the properties are compared with other signal functions. The sigmoidal function is monotonously increasing, continuous in all derivaties, and its output is 0.5 for zero input. The weight multiplication can be replaced by an addition when the training of the network is performed offline. We also present an approximation of this signal function, called a polygonal signal function, reducing the computational effort solely to bit sets and shift operations.

引用

页码：360 / 363

页数：3

共 11 条

[1] Double-exponential sigmoidal functions for neural networks
Elektrotech Informationstech E&I, 7-8 (360):
[2] DOUBLE-EXPONENTIAL UTILITY-FUNCTIONS
BELL, DE
MATHEMATICS OF OPERATIONS RESEARCH, 1986, 11 (02) : 351 - 361
[3] RESOLUTION-OPTIMAL EXPONENTIAL AND DOUBLE-EXPONENTIAL TRANSFORM METHODS FOR FUNCTIONS WITH ENDPOINT SINGULARITIES
Adcock, Ben
Martin-Vaquero, Jesus
Richardson, Mark
SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2017, 39 (01): : A164 - A187
[4] Double-exponential formula for infinite integrals of unilateral rapidly decreasing functions
Okayama, Tomoaki
JSIAM LETTERS, 2022, 14 : 17 - 20
[5] A modelling of standard current/voltage surges by means of one- and double-exponential functions
Kalat, Wlodzimierz
Archives of Electrical Engineering, 2009, 58 (227-228): : 71 - 80
[6] Neural networks for wind turbine supervisionNeurale Netze für die Windturbinenüberwachung
N. Elhor
R. Bertrand
J. G. Postaire
D. Hamad
e&i Elektrotechnik und Informationstechnik, 1999, 116 (6) : 366 - 369
[7] 15-digit accuracy calculations of Ambartsumian-Chandrasekhar's H-functions for four-term phase functions with the double-exponential formula
Kawabata, Kiyoshi
ASTROPHYSICS AND SPACE SCIENCE, 2018, 363 (01)
[8] Traversable wormhole modelling with exponential and hyperbolic shape functions in F(R, T) framework
Shweta
Mishra, Ambuj Kumar
Sharma, Umesh Kumar
INTERNATIONAL JOURNAL OF MODERN PHYSICS A, 2020, 35 (25):
[9] Confidence intervals prediction for the short-term electrical load neural forecasting modelsVorhersage des Konfidenzintervalls für neuronale modelle der kurzfristigen lastprognose
W. Bartkiewicz
e & i Elektrotechnik und Informationstechnik, 2000, 117 (1) : 8 - 12
[10] Development of a wind turbine gearbox virtual load sensor using multibody simulation and artificial neural networks; [Entwicklung eines virtuellen Lastsensors für Windturbinengetriebe mittels Mehrkörpersimulation und künstlicher neuronaler Netze]
Azzam B.
Schelenz R.
Roscher B.
Baseer A.
Jacobs G.
Forschung im Ingenieurwesen, 2021, 85 (2) : 241 - 250

← 1 2 →