Learning with Recurrent Neural Networks

• Main Author: Hammer, Barbara (Author, http://id.loc.gov/vocabulary/relators/aut)
• Corporate Author: SpringerLink (Online service)
• Format: Electronic eBook
• Language: English
• Published: London : Springer London : Imprint: Springer, 2000.
• Edition: 1st ed. 2000.
• Series: Lecture Notes in Control and Information Sciences, 254
• Subjects: Control engineering.; Robotics.; Mechatronics.; Control, Robotics, Mechatronics.

Folding networks, a generalisation of recurrent neural networks to tree structured inputs, are investigated as a mechanism to learn regularities on classical symbolic data, for example. The architecture, the training mechanism, and several applications in different areas are explained. Afterwards a...

Table of Contents:

• Introduction, Recurrent and Folding Networks: Definitions, Training, Background, Applications
• Approximation Ability: Foundationa, Approximation in Probability, Approximation in the Maximum Norm, Discussions and Open Questions
• Learnability: The Learning Scenario, PAC Learnability, Bounds on the VC-dimension of Folding Networks, Consquences for Learnability, Lower Bounds for the LRAAM, Discussion and Open Questions
• Complexity: The Loading Problem, The Perceptron Case, The Sigmoidal Case, Discussion and Open Questions
• Conclusion.