# Mathematical Perspectives on Neural Networks

By Paul Smolensky; Michael C. Mozer et al. | Go to book overview

9 Dynamical Systems

Morris W. Hirsch University of California at Berkeley

1. INTRODUCTION

In this chapter we look at neural nets as dynamical systems in the mathematical sense.

The following notations will be used. Rn denotes n-dimensional Euclidean space comprising all n-tuples of real numbers. We denote the inner or dot product of two vectors in Rn by x · y = 〈x, y〉 = Σ xiyi. The Euclidean norm of a vector x is

.

A map H : RnRm is smooth, or C1, provided it has continuous partial derivatives. In this case we denote the Jacobian matrix at p ∈ Rn by DH(p) = [(∂Hi/∂xj)(p)].We say H is C2 if the partial derivatives are C1. For integers k > 1 we recursively define H to be Ck if the partial derivatives are Ck-1. If f is a smooth real-valued function on Rn, the gradient vector grad f (p) of f at p ∈ Rn is the column vector of partial derivatives of f at p.

Before giving a formal mathematical treatment of dynamical systems, I indicate briefly how they arise in networks. Consider a net with n units labeled 1,. . ., n. For the moment assume that weights, thresholds, external inputs, and all other parameters are fixed. The activation of unit i at a given time is a real number, denoted here by ui. The state of the net is the activation vector u = (u1,. . . , un) ∈ Rn.

By starting at an initial activation state u and running the net, for each future time t > 0 we can determine the state x(t) to which state u will evolve at time t. Since we want to make a mathematical model of the net, we assume that, in principle, we can calculate any future state x(t) as a function of the time t and the initial state x(0) = u. The functions x(t), for all possible initial states u, constitute the activation dynamics of the net; they are called trajectories. To make explicit the fact that x(t) depends on u we use a notation such as x(t) = φt(u).

-271-

If you are trying to select text to create highlights or citations, remember that you must now click or tap on the first word, and then click or tap on the last word.
One moment ...
Default project is now your active project.
Project items
Notes

#### Cited page

Style
Citations are available only to our active members.
Buy instant access to cite pages or passages in MLA 8, MLA 7, APA and Chicago citation styles.

(Einhorn, 1992, p. 25)

(Einhorn 25)

(Einhorn 25)

1. Lois J. Einhorn, Abraham Lincoln, the Orator: Penetrating the Lincoln Legend (Westport, CT: Greenwood Press, 1992), 25, http://www.questia.com/read/27419298.

#### Cited page

Mathematical Perspectives on Neural Networks

Settings

#### Settings

Typeface
Text size Reset View mode
Search within

Look up

#### Look up a word

• Dictionary
• Thesaurus
Please submit a word or phrase above.

Why can't I print more than one page at a time?

Help
Full screen
Items saved from this book
• Bookmarks
• Highlights & Notes
• Citations
/ 864

### How to highlight and cite specific passages

1. Click or tap the first word you want to select.
2. Click or tap the last word you want to select, and you’ll see everything in between get selected.
3. You’ll then get a menu of options like creating a highlight or a citation from that passage of text.

## Cited passage

Style
Citations are available only to our active members.
Buy instant access to cite pages or passages in MLA 8, MLA 7, APA and Chicago citation styles.

"Portraying himself as an honest, ordinary person helped Lincoln identify with his audiences." (Einhorn, 1992, p. 25).

"Portraying himself as an honest, ordinary person helped Lincoln identify with his audiences." (Einhorn 25)

"Portraying himself as an honest, ordinary person helped Lincoln identify with his audiences." (Einhorn 25)

"Portraying himself as an honest, ordinary person helped Lincoln identify with his audiences."1

1. Lois J. Einhorn, Abraham Lincoln, the Orator: Penetrating the Lincoln Legend (Westport, CT: Greenwood Press, 1992), 25, http://www.questia.com/read/27419298.

## Thanks for trying Questia!

Please continue trying out our research tools, but please note, full functionality is available only to our active members.

Your work will be lost once you leave this Web page.