Zebra Tutor

Goal: Find consequences with minimal number of clues.

Clues? 

  • Natural language sentences
  • Logigram constraints

Idea:

How can we find consequences?

  • Reify theory rules
  • in a Theory that accepts prior knowledge (PK), a set of theory reification symbols and outputs (2-valued) reified knowledge

\( \exists clues :: \mathcal{Kb} : \exists reif : \exists prop, subj, val, dt: dt \neq U \land \forall extensions :: \mathcal{Kb} : Theory(PK, clues, reif, extensions) \Rightarrow extensions(prop,subj,val, dt)\)

Penne Farfall spirelli Arrabi Bologn Carbon
Andrea
Bart
Casper
arrabi
Bologn
Carbon
  • Andrea ate Penne
  • The person who ate penne, took Carbonara Sauce
  • Bart ate Penne or Farfalle
  • The spirelli was eaten with Bolognese

My instance

Penne Farfall spirelli Arrabi Bologn Carbon
Andrea X - - - - X
Bart - X - X - -
Casper - - X - X -
arrabi - X -
Bologn - - X
Carbon X - -
  • Andrea ate Penne
  • The person who ate penne, took Carbonara Sauce
  • Bart ate Penne or Farfalle
  • The spirelli was eaten with Bolognese

My instance

Transitivity

Zebra Tutor 2

By krr

Zebra Tutor 2

  • 1,238