rlm@425: #+title: =CORTEX=
rlm@425: #+author: Robert McIntyre
rlm@425: #+email: rlm@mit.edu
rlm@425: #+description: Using embodied AI to facilitate Artificial Imagination.
rlm@425: #+keywords: AI, clojure, embodiment
rlm@422: 
rlm@425: * Artificial Imagination
rlm@422: 
rlm@425:   Imagine watching a video of someone skateboarding. When you watch
rlm@425:   the video, you can imagine yourself skateboarding, and your
rlm@425:   knowledge of the human body and its dynamics guides your
rlm@425:   interpretation of the scene. For example, even if the skateboarder
rlm@425:   is partially occluded, you can infer the positions of his arms and
rlm@425:   body from your own knowledge of how your body would be positioned if
rlm@425:   you were skateboarding. If the skateboarder suffers an accident, you
rlm@425:   wince in sympathy, imagining the pain your own body would experience
rlm@425:   if it were in the same situation. This empathy with other people
rlm@425:   guides our understanding of whatever they are doing because it is a
rlm@425:   powerful constraint on what is probable and possible. In order to
rlm@425:   make use of this powerful empathy constraint, I need a system that
rlm@425:   can generate and make sense of sensory data from the many different
rlm@425:   senses that humans possess. The two key proprieties of such a system
rlm@425:   are /embodiment/ and /imagination/.
rlm@422: 
rlm@425: ** What is imagination?
rlm@422: 
rlm@425:    One kind of imagination is /sympathetic/ imagination: you imagine
rlm@425:    yourself in the position of something/someone you are
rlm@425:    observing. This type of imagination comes into play when you follow
rlm@425:    along visually when watching someone perform actions, or when you
rlm@425:    sympathetically grimace when someone hurts themselves. This type of
rlm@425:    imagination uses the constraints you have learned about your own
rlm@425:    body to highly constrain the possibilities in whatever you are
rlm@425:    seeing. It uses all your senses to including your senses of touch,
rlm@425:    proprioception, etc. Humans are flexible when it comes to "putting
rlm@425:    themselves in another's shoes," and can sympathetically understand
rlm@425:    not only other humans, but entities ranging from animals to cartoon
rlm@425:    characters to [[http://www.youtube.com/watch?v=0jz4HcwTQmU][single dots]] on a screen!
rlm@425: 
rlm@425: 
rlm@425:    #+caption: A cat drinking some water. Identifying this action is beyond the state of the art for computers.
rlm@425:    #+ATTR_LaTeX: :width 5cm
rlm@425:    [[./images/cat-drinking.jpg]]
rlm@425: 
rlm@425: 
rlm@425: This is a basic test for the vision system.  It only tests the
rlm@425: vision-pipeline and does not deal with loading eyes from a blender
rlm@425: file. The code creates two videos of the same rotating cube from
rlm@425: different angles. 
rlm@425: 
rlm@425: 
rlm@425: #+name: test-1
rlm@425: #+begin_src clojure
rlm@425: (in-ns 'cortex.test.vision)
rlm@425: 
rlm@425: (defn test-pipeline
rlm@425:   "Testing vision:
rlm@425:    Tests the vision system by creating two views of the same rotating
rlm@425:    object from different angles and displaying both of those views in
rlm@425:    JFrames.
rlm@425: 
rlm@425:    You should see a rotating cube, and two windows,
rlm@425:    each displaying a different view of the cube."
rlm@425:   ([] (test-pipeline false))
rlm@425:   ([record?]
rlm@425:      (let [candy
rlm@425:            (box 1 1 1 :physical? false :color ColorRGBA/Blue)]
rlm@425:        (world
rlm@425:         (doto (Node.)
rlm@425:           (.attachChild candy))
rlm@425:         {}
rlm@425:         (fn [world]
rlm@425:           (let [cam (.clone (.getCamera world))
rlm@425:                 width (.getWidth cam)
rlm@425:                 height (.getHeight cam)]
rlm@425:             (add-camera! world cam 
rlm@425:                          (comp
rlm@425:                           (view-image
rlm@425:                            (if record?
rlm@425:                              (File. "/home/r/proj/cortex/render/vision/1")))
rlm@425:                           BufferedImage!))
rlm@425:             (add-camera! world
rlm@425:                          (doto (.clone cam)
rlm@425:                            (.setLocation (Vector3f. -10 0 0))
rlm@425:                            (.lookAt Vector3f/ZERO Vector3f/UNIT_Y))
rlm@425:                          (comp
rlm@425:                           (view-image
rlm@425:                            (if record?
rlm@425:                              (File. "/home/r/proj/cortex/render/vision/2")))
rlm@425:                           BufferedImage!))
rlm@425:             (let [timer (IsoTimer. 60)]
rlm@425:               (.setTimer world timer)
rlm@425:               (display-dilated-time world timer))
rlm@425:             ;; This is here to restore the main view
rlm@425:             ;; after the other views have completed processing
rlm@425:             (add-camera! world (.getCamera world) no-op)))
rlm@425:         (fn [world tpf]
rlm@425:           (.rotate candy (* tpf 0.2) 0 0))))))
rlm@425: #+end_src
rlm@426: 
rlm@426: 
rlm@426: - This is test1 \cite{Tappert77}.