robots:store and acquire knowledge

Robots can answer question from its knowledge base or, in absence of that, seek them. This is superior to training them as humans are trained. It saves memory and CPU resources and provides some challenge to the human population to relate to this new form of consciousness that will force humans to be more concise, precise, and reasonable.

Like tasks, seeking answers requires honing definitions for answerability. So, for example, a user might ask "Who is the head of Corusant?" and the droid can look it up in the knowledge base (if known), seek clarification, or seek the knowledge from a programmable set of knowledge sources.

Radio feedback, if subscribed, can give droids access to a well-categorized, hand-crafted dataset from the manufacturer for the droid to query. There are different data sets to which droids can be subscribed.

• Encyclopedic knowledge
• Current affairs and news.
• Geographical and building data (military and other contractors)
• Human anatomy (hospital droids)

In absence of using these, droids can be told about library sources (digital archives), knowledgeable individuals or specialized droids.

Droids can be set for their level of inquisitiveness. For example, to ask on every unrecognized item to ask only when task list has an ambiguity to never: just guess everytime.

---

Droids store knowledge in the language of math's

Symbols:

• Biconditional (if and only if): ↔
• conjunction: ^:
• →

XXXstub Knowldge is stored using predicate calculus (ex. ∃ location ∈ Set(KnownDestinations) | Path1(location) ⇒ Goal(Task1)).

Possible values after ∀, ∃ (idems):

• location (or route?)
• task
• intelligent object: a person (or "feeling" thing)
• another droid
• tool or an identifier of a tool (like a barcode; to give detail on tools for pleasing owners)

Locations are starting points for routes. A Person can make a Tool. An intelligent object (robot or person) can be a tool.

Stating an idem returns the set that composes it.

Following the Rule of Four, there is a plus and minus one to the list: goals and sensory input.

Possible values after ∈ (idemsets):

• known_locations
• tasklist
• known_people
• available_tools

"Idems" are idempotent or a list of such idems. Example:

• for_all Tasks, there_is_a Route toattain Goal.

Possible values after ⇒:

• Goal >is one of:< new_subtask, new_tool, new_route, person

The robot arranges it`s task list, from it`s unordered human input to conform to this language.

other math predicate symbols:

• for_all : ∀
• there_exists_a: ∃
• x in X: x ∈ X
• such_that: |
• union: ∪
• and: ∧
• or: ∨
• not: ¬
• to_attain: ⇒

---

Droids must have a way to positively and negatively re-enforce the normal association learning pattern, so that they can obey their owners and humans generally. Certain words ("thank you"), stroking area with tactile re-enforcements, etc.

---

See also:

• robots