Mercurial > cortex

comparison thesis/abstract.org @ 432:1e5ea711857d

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
abstract first draft.
author Robert McIntyre <rlm@mit.edu>
date Sun, 23 Mar 2014 16:33:01 -0400
parents
children d52bff980f0d
comparison
equal deleted inserted replaced
431:7410f0d8011c 432:1e5ea711857d
1 Here I explore the design and capabilities of my system (called
2 =CORTEX=) which enables experiments in /embodied artificial
3 intelligence/ -- that is, AI which uses a physical simulation of
4 reality accompanied by a simulated body to solve problems.
5
6 In the first half of the thesis I describe the construction of
7 =CORTEX= and the rationale behind my architecture choices. =CORTEX= is
8 a complete platform for embodied AI research. It provides multiple
9 senses for simulated creatures, including vision, touch,
10 proprioception, muscle tension, and hearing. Each of these senses
11 provides a wealth of parameters that are biologically
12 inspired. =CORTEX= is able to simulate any number of creatures and
13 senses, and provides facilities for easily modeling and creating new
14 creatures. As a research platform it is more complete than any other
15 system currently available.
16
17 In the second half of the thesis I develop a computational model of
18 empathy, using =CORTEX= as a base. Empathy in this context is the
19 ability to observe another creature and infer what sorts of sensations
20 that creature is feeling. My empathy algorithm involves multiple
21 phases. First is free-play, where the creature moves around and gains
22 sensory experience. From this experience I construct a representation
23 of the creature's sensory state space, which I call \phi-space. Using
24 \phi-space, I construct an efficient function for enriching the
25 limited data that comes from observing another creature with a full
26 compliment of imagined sensory data based on previous experience. I
27 can then use the imagined sensory data to recognize what the observed
28 creature is doing and feeling, using straightforward embodied action
29 predicates. This is all demonstrated with using a simple worm-like
30 creature, recognizing worm-actions in video.
31