view thesis/dylan-accept.diff @ 525:25f23cfd56ce

alter pics, enhance text.
author Robert McIntyre <rlm@mit.edu>
date Mon, 21 Apr 2014 02:13:23 -0400
parents 447c3c8405a2
children
line wrap: on
line source
1 @@ -3210,13 +3329,14 @@
3 In this thesis you have seen the =CORTEX= system, a complete
4 environment for creating simulated creatures. You have seen how to
5 - implement five senses including touch, proprioception, hearing,
6 - vision, and muscle tension. You have seen how to create new creatues
7 - using blender, a 3D modeling tool. I hope that =CORTEX= will be
8 - useful in further research projects. To this end I have included the
9 - full source to =CORTEX= along with a large suite of tests and
10 - examples. I have also created a user guide for =CORTEX= which is
11 - inculded in an appendix to this thesis.
12 + implement five senses: touch, proprioception, hearing, vision, and
13 + muscle tension. You have seen how to create new creatues using
14 + blender, a 3D modeling tool. I hope that =CORTEX= will be useful in
15 + further research projects. To this end I have included the full
16 + source to =CORTEX= along with a large suite of tests and examples. I
17 + have also created a user guide for =CORTEX= which is inculded in an
18 + appendix to this thesis \ref{}.
19 +# dxh: todo reference appendix
21 You have also seen how I used =CORTEX= as a platform to attach the
22 /action recognition/ problem, which is the problem of recognizing
24 @@ -2888,7 +3007,7 @@
25 #+end_src
26 #+end_listing
28 -** Efficient action recognition with =EMPATH=
29 +** =EMPATH= recognizes actions efficiently
31 To use =EMPATH= with the worm, I first need to gather a set of
32 experiences from the worm that includes the actions I want to
34 @@ -3044,9 +3163,9 @@
35 to interpretation, and dissaggrement between empathy and experience
36 is more excusable.
38 -** Digression: bootstrapping touch using free exploration
39 -
40 - In the previous section I showed how to compute actions in terms of
41 +** Digression: Learn touch sensor layout through haptic experimentation, instead
42 +# Boostraping touch using free exploration
43 +In the previous section I showed how to compute actions in terms of
44 body-centered predicates which relied averate touch activation of
45 pre-defined regions of the worm's skin. What if, instead of recieving
46 touch pre-grouped into the six faces of each worm segment, the true
48 @@ -2193,7 +2311,7 @@
49 #+ATTR_LaTeX: :width 11cm
50 [[./images/proprio.png]]
52 -** Muscles are both effectors and sensors
53 +** Muscles contain both sensors and effectors
55 Surprisingly enough, terrestrial creatures only move by using
56 torque applied about their joints. There's not a single straight
58 @@ -2059,7 +2177,7 @@
59 #+ATTR_LaTeX: :width 15cm
60 [[./images/touch-cube.png]]
62 -** Proprioception is the sense that makes everything ``real''
63 +** Proprioception provides knowledge of your own body's position
65 Close your eyes, and touch your nose with your right index finger.
66 How did you do it? You could not see your hand, and neither your
68 @@ -1257,8 +1375,8 @@
69 community and is now (in modified form) part of a system for
70 capturing in-game video to a file.
72 -** Hearing is hard; =CORTEX= does it right
73 -
74 +** ...but hearing must be built from scratch
75 +# is hard; =CORTEX= does it right
76 At the end of this section I will have simulated ears that work the
77 same way as the simulated eyes in the last section. I will be able to
78 place any number of ear-nodes in a blender file, and they will bind to
80 @@ -1565,7 +1683,7 @@
81 jMonkeyEngine3 community and is used to record audio for demo
82 videos.
84 -** Touch uses hundreds of hair-like elements
85 +** Hundreds of hair-like elements provide a sense of touch
87 Touch is critical to navigation and spatial reasoning and as such I
88 need a simulated version of it to give to my AI creatures.
90 @@ -956,7 +1074,7 @@
91 #+ATTR_LaTeX: :width 15cm
92 [[./images/physical-hand.png]]
94 -** Eyes reuse standard video game components
95 +** Sight reuses standard video game components...
97 Vision is one of the most important senses for humans, so I need to
98 build a simulated sense of vision for my AI. I will do this with
99 @@ -459,8 +577,8 @@
100 simulations of very simple creatures in =CORTEX= generally run at
101 40x on my machine!
103 -** What is a sense?
104 -
105 +** All sense organs are two-dimensional surfaces
106 +# What is a sense?
107 If =CORTEX= is to support a wide variety of senses, it would help
108 to have a better understanding of what a ``sense'' actually is!
109 While vision, touch, and hearing all seem like they are quite