Natural Gradient Descent for Training Stochastic Complex-Valued Neural Networks

Abstract

In this paper, the natural gradient descent method for the multilayer stochastic complex-valued neural networks is considered, and the natural gradient is given for a single stochastic complex-valued neuron as an example. Since the space of the learnable parameters of stochastic complex-valued neural networks is not the Euclidean space but a curved manifold, the complex-valued natural gradient method is expected to exhibit excellent learning performance.

Authors and Affiliations

Tohru Nitta

Keywords

Related Articles

Crowdsensing: Socio-Technical Challenges and Opportunities

With the advancement in mobile technology, the sensing and computational capability of mobile devices is increasing. The sensors in mobile devices are being used in a variety of ways to sense and actuate. Mobile crowdsen...

Markovian Process and Novel Secure Algorithm for Big Data in Two-Hop Wireless Networks

This paper checks the correctness of our novel algorithm for secure, reliable and flexible transmission of big data in two-hop wireless networks using cooperative jamming scheme of attacker location unknown through Marko...

KNN and ANN-based Recognition of Handwritten Pashto Letters using Zoning Features

This paper presents an intelligent recognition sys-tem for handwritten Pashto letters. However, handwritten char-acter recognition is challenging due to the variations in shape and style. In addition to that, these chara...

How Volunteering Affects the Offender’s Behavior

Agent Based modelling is widely used for presenting and evaluating a social phenomenon. Agent based modelling helps the researcher to analyze and evaluate a social model and its related hypothetical theories by simulatin...

Dual-Cross-Polarized Antenna Decoupling for 43 GHz Planar Massive MIMO in Full Duplex Single Channel Communications

Massive Multiple Input Multiple Output (MIMO) and Full Duplex Single Channel (FDSC) at mm-Wave are key technology of future advanced wireless communications. Self-interference is the main problem in this technique becaus...

Download PDF file
  • EP ID EP121478
  • DOI 10.14569/IJACSA.2014.050729
  • Views 101
  • Downloads 0

How To Cite

Tohru Nitta (2014). Natural Gradient Descent for Training Stochastic Complex-Valued Neural Networks. International Journal of Advanced Computer Science & Applications, 5(7), 193-198. https://europub.co.uk/articles/-A-121478