Switches to using proper figures for all numbered figures.

Closes #9

machine_in_ghost.tex

@@ -7067,47 +7067,23 @@ \chapter{A Human's Guide to Words}
 
 {
  Then I might design a neural network that looks something like
-Figure 161.1.}
+Figure \ref{nc_network_1}.}
 
-{
- ~}
-
-{\centering
-\mygraphics{images/img198.jpg}
- \newline
- Figure 161.1: Network 1
-\par}
-
-
-\bigskip
-
-{
- ~}
+%161.1
+\myfigurec{images/img198.jpg}{nc_network_1}{Network 1}
 
 {
  Network 1 is for classifying bleggs and rubes. But since
 ``blegg'' is an unfamiliar and
 synthetic concept, I've also included a similar Network
-1b in Figure 161.2 for distinguishing humans from Space Monsters, with
+1b in Figure \ref{nc_network_1b} for distinguishing humans from Space Monsters, with
 input from Aristotle (``All men are
 mortal'') and Plato's Academy
 (``A featherless biped with broad
 nails'').}
 
-{
- ~}
-
-{\centering
-\mygraphics{images/img199.jpg}
- \newline
- Figure 161.2: Network 1b
-\par}
-
-
-\bigskip
-
-{
- ~}
+%161.2
+\myfigurec{images/img199.jpg}{nc_network_1b}{Network 1b}
 
 {
  A neural network needs a learning rule. The obvious idea is that
@@ -7187,23 +7163,11 @@ \chapter{A Human's Guide to Words}
 {
  So what might be a more realistic neural network design?}
 
-{
- ~}
-
-{\centering
-\mygraphics{images/img200.jpg}
- \newline
- Figure 161.3: Network 2
-\par}
-
-
-\bigskip
-
-{
- ~}
+%161.3
+\myfigurec{images/img200.jpg}{nc_network_2}{Network 2}
 
 {
- In Network 2 of Figure 161.3, a wave of activation converges on
+ In Network 2 of Figure \ref{nc_network_2}, a wave of activation converges on
 the central node from any clamped (observed) nodes, and then surges
 back out again to any unclamped (unobserved) nodes. Which means we can
 compute the answer in one step, rather than waiting for the network to
@@ -7387,37 +7351,17 @@ \chapter{A Human's Guide to Words}
  So why might someone feel an impulse to go on arguing whether the
 object is \textit{really} a blegg?}
 
-{
- ~}
-
-{\centering
-\mygraphics{images/img198.jpg}
- \newline
- Figure 162.1: Network 1
-\par}
-
-
-\bigskip
-
-{
- ~}
-
-{\centering
-\mygraphics{images/img200.jpg}
- \newline
- Figure 162.2: Network 2
-\par}
-
+%162.1
+\myfigurec{images/img198.jpg}{afi_network_1}{Network 1}
 
-\bigskip
+%162.2
+\myfigurec{images/img200.jpg}{afi_network_2}{Network 2}
 
-{
- ~}
 
 {
  These diagrams from Neural Categories show two different neural
 networks that might be used to answer questions about bleggs and rubes.
-Network 1 (Figure 162.1) has a number of disadvantages---such as
+Network 1 (Figure \ref{afi_network_1}) has a number of disadvantages---such as
 potentially oscillating/chaotic behavior, or requiring
 $O(N^{2})$ connections---but Network 1's
 structure does have one major advantage over Network 2: every unit in
@@ -7426,7 +7370,7 @@ \chapter{A Human's Guide to Words}
 left over.}
 
 {
- Network 2 (Figure 162.2), however, is a far better candidate for
+ Network 2 (Figure \ref{afi_network_2}), however, is a far better candidate for
 being something vaguely like how the human brain works:
 It's fast, cheap, scalable---and has an extra dangling
 unit in the center, whose activation can still vary, even after
@@ -7904,20 +7848,8 @@ \chapter{A Human's Guide to Words}
 they categorize as \textit{aiiieeee} CRUNCH CRUNCH \textit{help
 it's got my arm} CRUNCH GULP.''}
 
-{
- ~}
-
-{\centering
-\mygraphics{images/img206.jpg}
- \newline
- Figure 164.1: Network 3
-\par}
-
-
-\bigskip
-
-{
- ~}
+%164.1
+\myfigurec{images/img206.jpg}{fm_network_3}{Network 3}
 
 {
  Considering this as a design constraint on the human cognitive
@@ -10933,22 +10865,10 @@ \chapter{A Human's Guide to Words}
 in $U$,$V$, $W$, $X$, and $Y$.}
 
 {
- Sound familiar? It should; see Figure 177.1.}
-
-{
- ~}
-
-{\centering
-\mygraphics{images/img200.jpg}
- \newline
- Figure 177.1: Network 2
-\par}
+ Sound familiar? It should; see Figure \ref{cinb_network_2}.}
 
-
-\bigskip
-
-{
- ~}
+%177.1
+\myfigurec{images/img200.jpg}{cinb_network_2}{Network 2}
 
 {
  As a matter of fact, if you use the right kind of neural network