Abductive Constraint Logic Programming

A single-file abduction system for prolog. Doesn't need/use CHR for abduction.

Tested with SWI, Trealla, Scryer, and Tau:

?- [alp_int].
Warning: singleton: RestOfLiterals, near alp_int:160
Warning: singleton: RestOfLiterals, near alp_int:169
Warning: singleton: Y, near alp_int:177
Warning: singleton: Z, near alp_int:178
   true.
?- [alp_prof_ex].
   true.
?- demo([professor(henning)], [], FinalExp).
   FinalExp = [not(rich(henning)),professor(henning)]
;  false.
?- demo([happy(henning)], [], FinalExp).
   FinalExp = [not(professor(henning)),rich(henning)]
;  FinalExp = [has(henning,nice_students),not(rich(henning)),professor(henning)]
;  false.

See more examples in alp_exp.pl.

• Integrity constraints (ic) tell the abductive system what can't be true. For example, in alp_prof_ex, you can't have a rich professor.
• Abducibles tell the abductive system how to think about facts that it's seeing for the first time (open world). For example, in EX3 of alp_exp, if you ask about a new bird that isn't tweety or sam, then the system will say that this new bird is either slim or bird_ab. Abducibles represent default logic.

• Website
• Abduction and language processing with CHR