D etecti ng and Correcti ng Errors of O m i ssi on A fter Expl anati on-based Learni ng M i chael J. Pazzani Depar tm ent of I nf or m ati on and Com puter Sci ence Uni versi ty of Cal i forni a, I rvi ne, CA 92717 pazzani @ i cs. uci . edu 1 I JCA I - 89 M onday, A ugust 14, 1989
O utl i ne I .Probl em : A . D etecti ng er r or s i n gener al i zati on pr oduced by EBL B. A ssi gni ng bl am e to r ul es i n dom ai n theor y C. Cor r ecti ng dom ai n theor y I I . Types of Errors A . Errors of O m i ssi on- Fai l to m ake cor r ect pr edi cti on B. Errors of Com m i ssi on - M ake i ncorrect predi cti on I I I . I ndexi ng gener al i zati ons i n M em or y A . Expl anatory- Gi ven resul t, Predi ct cause B. Predi cti ve- Gi ven acti on, Pr edi ct r esul t I V . Unsuper vi sed detecti on of er r or s of om i ssi on V . Bl am e A ssi gnm ent & r evi si ng dom ai n theor y V I . Experi m ental Resul ts 2 I JCA I - 89 M onday, A ugust 14, 1989
Backgr ound: O CCA M • Perform ance task • Pr edi ct outcom e of econom i c sancti on i nci dents • O utcom e i nferred by hi erarchi cal cl assi fi cati on • Er r or of om i ssi on occur s i f i nci dent cannot be cl assi fi ed • Error of com m i ssi on occurs i f i nci dent i s cl assi fi ed i ncorrectl y coerce s1 s2 s3 s7 s5 s6 s8 • Lear ni ng M ethod: •Com bi nes em pi r i cal and expl anati on- based l ear ni ng • em pi r i cal techni ques l ear n the dom ai n theor y used by EBL → • Probl em : i ncor r ect dom ai n theor y i ncorrect general i zati ons 3 I JCA I - 89 M onday, A ugust 14, 1989
Term i nol ogy Types of exam pl es: Foundati onal : exam pl es f r om w hi ch the dom ai n theor y i s l ear ned E. g. , Par ent hel pi ng chi l d Perform ance: exam pl es of the per f or m ance task E. g. , Ki dnappi ng exam pl es Foundati onal exam pl es ar e subpr obl em s of the per f or m ance task. Per f or m ance exam pl es ar e exam pl es of the per f or m ance task. D i sti ncti on i s r el ati ve to task. I f task i s pr edi cti ng w hom to sel l r ansom i nsur ance to, the ki dnappi ng exam pl es ar e f oundati onal . Types of r ul es (or schem ata) Dom ai n: used by EBL to expl ai n per f or m ance exam pl e Com pi l ed: cr eated by EBL to be used i n per f or m ance task. 4 I JCA I - 89 M onday, A ugust 14, 1989
Pr obl em Statem ent: • Rul e i n dom ai n theor y ar e l ear ned (or hand- coded) Par ents have a goal of pr eser vi ng thei r chi l dr en’ s heal th • Cr eate com pi l ed r ul e w i th EBL O ne pl an to obtai n m oney i s to thr eaten to ki l l the chi l d of a ri ch person. • D etect er r or of om i ssi on i n com pi l ed r ul e fr om per for m ance exam pl e A ki dnapper obtai ns m oney f r om gr andpar ent of hostage. • A ssi gn Bl am e for error on rul e i n dom ai n theory • Revi se r ul e i n dom ai n theor y M em ber s of the sam e f am i l y have a goal of pr eser vi ng each other ’ s heal th • Revi se Rul e i n com pi l ed theor y 5 I JCA I - 89 M onday, A ugust 14, 1989
W hy Errors of O m i ssi on? O nl y type of error created by one- si ded l earni ng al gori thm s • I ncr em ental , hi l l - cl i m bi ng al gor i thm s (Langl ey et al . , 1987) . • A ccounts f or som e hum an l ear ni ng • Gram m ar: (Berw i ck, 1986) • Concept A cqui si ti on (Br uner , et al , 1956) • Subj ect of theor eti cal anal ysi s (V al i ant, 84; H aussl er , 87) • H ypothesi s never m ore general than “true” hypothesi s. I ncr em ental bl am e assi gnm ent and r evi si on possi bl e: Error of O m i ssi on: O ne of the dom ai n theor y r ul es used to cr eate a com pi l ed r ul e needs to be gener al i zed by dr oppi ng a condi ti on that i s not pr esent i n a per f or m ance exam pl e Error of com m i ssi on: O ne of the dom ai n theor y r ul es used to cr eate a com pi l ed r ul e needs to be speci al i zed by addi ng a condi ti on that i s pr esent i n a new per f or m ance exam pl e. 6 I JCA I - 89 M onday, A ugust 14, 1989
Unsuper vi sed detecti on of er r or of om i ssi on coerce s1 s2 s3 s7 s5 s6 s8 Probl em : Fi nd a schem a i n m em or y that w oul d pr edi ct the outcom e of a new per f or m ance exam pl e, i f the schem a w er e m or e gener al . A ppr oach: Di sti ngui sh betw een tw o uses of schem ata Predi cti ve: What would happen if the United States refused to sell computers to South Korea unless South Korea stopped exporting automobiles to Canada? Expl anatory: What could cause the price of oil to rise? 7 I JCA I - 89 M onday, A ugust 14, 1989
D er i vati on and use of I ndi ces D eri ved anal yti cal l y: expl anator y: Featur es f r om consequent of dom ai n r ul es. pr edi cti ve: Featur es f r om antecedent of dom ai n r ul es. Use duri ng retri eval : expl anator y: Fi ndi ng a schem a to expl ai n the cause of an outcom e predi cti ve: Fi ndi ng a schem a to pr edi ct the outcom e of an event. D eter m i ni ng i ndi ces: A n exam pl e A ustr al i a and Fr ance, 1983 I n 1983, A ustr al i a r efused to sel l ur ani um to Fr ance, unl ess Fr ance ceased nucl ear testi ng i n the South Paci f i c. Fr ance pai d a hi gher pr i ce to buy ur ani um f r om South A f r i ca and conti nued nucl ear testi ng. Pr ocessi ng goal : Expl ai n r esul t (Fr ance buys ur ani um el sew her e) 8 I JCA I - 89 M onday, A ugust 14, 1989
D eter m i ni ng i ndi ces: A n exam pl e 1. Thr eat - > I ncr eased dem and (ACT TYPE (SELL) ACTOR (POLITY EXPORTS ?Y ECONOMY (FREE)) TO ?X:(POLITY IMPORTS ?Y ECONOMY (FREE)) OBJECT ?Y:(COMMODITY) MODE (NEG)) result (STATE TYPE (DEMAND-INCREASE) ACTOR ?X OBJECT ?Y) 2. I ncr eased dem and - > W i l l i ngness to pay hi gher pr i ce 3. Purchase - > Possess 9 I JCA I - 89 M onday, A ugust 14, 1989
I ndexi ng by pr edi cti ve and expl anator y f eatur es A B C D coerce actor target ... response outcome (coerce actor (polity exports =OBJECT economy (free)) target (polity economic-health (strong) economy (free) imports =OBJECT) ... response (act type (sell) actor (polity bus-rel =TARGET exports =OBJECT) object =OBJECT price (money value (>market)) to =TARGET) outcome (goal-outcome type (failure)) 10 I JCA I - 89 M onday, A ugust 14, 1989
D etecti ng an er r or of om i ssi on i nput: N ew per f or m ance exam pl e Hi er ar chy of schem a (dom ai n & com pi l ed theor y) Retr i eve schem a by fol l ow i ng pr edi cti ve i ndi ces I f schem a has outcom e , Then I f outcom e of exam pl e and schem a agr ee Then EXI T El se “Error of com m i ssi on” El se I f exam pl e i s expl ai nabl e Then EBL(exam pl e) El se I f r etr i eve schem a by expl anator y i ndi ces Then “attem pt bl am e assi gnm ent” El se TDL(exam pl e) or SBL(exam pl e) Fi nd a schem a suffi ci entl y cl ose to the per f or m ance exam pl e that w oul d expl ai n the exam pl e i f gener al i zed. 11 I JCA I - 89 M onday, A ugust 14, 1989
Bl am e assi gnm ent (defi ni ng “suffi ci entl y cl ose”) 1. Reexpl ai n the schem a w i th EBL that m ai ntai ns dependenci es betw een constr ai nts i n the com pi l ed r ul e and r ul es i n the dom ai n theor y. (Tr ade- of f betw een stor i ng & r ecom puti ng): (coerce target (polity economic-health (strong) <- Rule.013 economy (free) <- Rule.012 imports =OBJECT) <- {Rule.01 Rule.13} 2. Fi nd di f f er ences betw een the schem a and the new event. 3. Col l ect i nference rul es responsi bl e for di fferences. 4. I f one i nference rul e i s responsi bl e for al l the di fferences, then assi gn bl am e to thi s i nf er ence r ul e. 12 I JCA I - 89 M onday, A ugust 14, 1989
Recommend
More recommend
