An Architecture for Knowledge Based Deduction

Th i s paper d e s c r i b e s the d e s i g n o f a g e n e r a l purpose d e d u c t i o n eng ine f o r use i n expe r t sys tems. I t r e p r e s e n t s an e x t e n s i o n o f a n a t u r a l d e d u c t i o n theorem p r o v e r t h a t has e l i m i n a t e d the n e g a t i o n symbol in f a v o r o f c e r t a i n t y f a c t o r s and has an improved method o f s p l i t t i n g problems i n t o subp rob lems . Formulas a re o b j e c t c e n t e r e d and are s t o r e d a t nodes w h i c h a re des igned to c o n t r o l and p r o v i d e sha rpe r f ocus to t h e s e a r c h . These nodes e i t h e r a re p o i n t e d to by words in a d i c t i o n a r y o r a re c r e a t e d b y t he d e d u c t i o n eng ine i n i t s a t t emp t t o s o l v e a s p e c i f i c p r o b l e m . 1 . I n t r o d u c t i o n T h i s paper d e s c r i b e s t he d e s i g n o f a g e n e r a l purpose d e d u c t i o n eng ine f o r use i n e x p e r t sys tems. I t goes beyond MYCIN [ 8 ] , [9 ] i n u t i l i z i n g more o f t h e power o f p r e d i c a t e l o g i c , w h i l e a t the same t i m e , i t s t r e a t m e n t o f c e r t a i n t y f a c t o r s i s no t i n t e r t w i n e d w i t h a p a r t i c u l a r sea rch s t r a t e g y (such as the backward c h a i n i n g g a t h e r i n g o f ev idence f o r c e r t a i n t y f a c t o r e v a l u a t i o n used b y MYCIN). The p r e s e n t system employs n a t u r a l d e d u c t i o n [1] , [ 3 ] , [A] wh i ch o f f e r s some advantages over r e s o l u t i o n [7] i n t h a t 1 ) i t does no t c o n v e r t t o c l a u s e f o r m , t h e r e b y making i t s r e p r e s e n t a t i o n c l o s e r to t h a t used by humans and s a v i n g p o t e n t i a l l y h e u r i s t i c i n f o r m a t i o n , and 2 ) i t i s e a s i e r t o s p l i t a p rob lem i n t o subproblems s i n c e e x p r e s s i o n s d o no t ge t " m u l t i p l i e d o u t " d u r i n g the c o n v e r s i o n t o c l a u s e f o r m . The system m a i n t a i n s a d i c t i o n a r y o f wo rds , each word p o i n t i n g to one o r more s u b j e c t a r e a s , w i t h each s u b j e c t a rea c o n t a i n i n g f o rmu las r e l e v a n t t o t he s u b j e c t . T h i s enab les i t t o p rocess a l a r g e number of f o r m u l a s by f o c u s i n g on s u b j e c t areas r e l e v a n t to t h e q u e r y . The system has a d e d u c t i o n module whose p r i m a r y purpose i s to o b t a i n answers t o q u e r i e s and w i l l have a knowledge a c q u i s i t i o n module whose purpose is to add new fo rmu las to the d a t a base and upda te the d i c t i o n a r y . Th i s paper i s concerned p r i m a r i l y w i t h t h e d e d u c t i o n modu le . T h i s r e s e a r c h was s u p p o r t e d in p a r t by the N a t i o n a l Sc ience F o u n d a t i o n under Grant DCR-8408389.