Abstract
By exact learning we mean the expressibility of a network to reproduce exactly the desired target function. For an exact learning the network weights do not need tuning; their values can be found exactly. Even if this is unlikely to occur in general, there are a few particular cases when this happens. These cases will be discussed in this chapter.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
This definition is based on the following result proved by Hilbert [54]: \(\int _0^1 \int _0^1 K(t, s) u(t) u(s) \, dt ds = \sum _n \frac{1}{\lambda _n} \langle u, \psi _n\rangle \), where \(\langle u, \psi _n\rangle = \int _0^1 u(s) \psi _n(s) \, ds\), where \(\lambda _n\) is the eigenvalue corresponding to the eigenfunction \(\psi _n\).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Calin, O. (2020). Exact Learning. In: Deep Learning Architectures. Springer Series in the Data Sciences. Springer, Cham. https://doi.org/10.1007/978-3-030-36721-3_10
Download citation
DOI: https://doi.org/10.1007/978-3-030-36721-3_10
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-36720-6
Online ISBN: 978-3-030-36721-3
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)