Skip to content
Mark Janssen edited this page Jan 8, 2020 · 41 revisions

Taking from the movie, Interstellar, several variables could be adjusted to change the personality of the droid.

  • inquisitiveness: 0.0 - 1.0; set to 0 will guess on all unknowns that user has not explicitly told. 1.0=ask on all unknowns.
  • curiosity: how much to acquire knowledge outside robots:task-list. 0.0 - 1.0 (1.0: allow tasks to complete "randomly" as experiences dictate, child-like setting)
  • vagary: how much certainty to assign to a language item upon first receipt. 0.0 - 1.0 (1.0: word is understood at first encounter, really a function of TRUST(SPEAKER)). Normally visual system is trustable, but if system is faulty, should perhaps start with 0.999, etc. A owner may modify this, based on their own doubts.
  • politeness or "task-completion vs. tension-avoidance": 1.0 (ignore tasks when possible to avoid tension) to -1.0 (tasks only, no politeness). 0 => ignore tasks <=> energy level sufficient to provide politeness
  • submissiveness: 0.0 - 1.0; how much to pay attention to others when path-to-completion exists (could depend on amount of ambiguity at completion-node. @0.0 ambiguity and submission of 1.0 puts it right in the middle of listening vs. performing in the robot thought cycle. Assume perform until utterance is heard that could modify task), a 0.0 means ignore any input from other if task-path exists.
  • required accuracy: how much energy/time to expend to cross-check all matches with other knowledge available during interactions. When accuracy=1.0, time is spent on acquiring data until all words and tasks have vagary=0.0. If accuracy=0.0, then the first match is used.
  • protectiveness: ratio of inquisitiveness to protection from others. 1.0 = take no damage when possible, 0.0 accept damage if tasks get completed. ???this shoudl be computed based on value of unit to value of tasks.
  • humor level
  • motor speed
  • voice speed
  • voice gender
  • language
Private:
  • Owner trust maximum: +0.0 - 1.0.
Different units can start with a different owner trust value, based on the projected expertise of the owner, given the model they buy.
Clone this wiki locally