Tables
|
|
|
³í¸® ¿¬»êÀÚ
( ¶Ç´Â connectives )ÀÇ Á¾·ù
ºÎÁ¤ (negation) |
~P, NOT P, P`, ¡þP |
³í¸®°ö (conjuction) |
P¡üQ, P AND Q, P*Q, P&Q µÎ¸íÁ¦°¡ ¸ðµÎ ÂüÀÏ ¶§¸¸ Âü |
³í¸®ÇÕ (disjunction) |
P¡ýQ, P OR Q, P+Q, µÎ¸íÁ¦°¡ ¸ðµÎ °ÅÁþÀÏ ¶§¸¸ °ÅÁþ |
¹èŸÀû³í¸®ÇÕ (exclusive disjunction) |
P XOR Q , ¸ðµÎ ÂüÀ̰ųª °ÅÁþÀÌ¸é °ÅÁþ |
Á¶°Ç (conditional) ¶Ç´Â implication |
P->Q '(¸¸ÀÏ) PÀ̸é QÀÌ´Ù' P°¡ ÂüÀ̰í Q°¡ °ÅÁþÀÏ ¶§¸¸ °ÅÁþ |
½ÖÁ¶°Ç (biconditional) |
P<->Q , (P->Q)¡ü(Q->P)ÀÎ °æ¿ì °°Àº Áø¸®°ªÀ» °¡Áú ¶§¸¸ Âü |
Deduction |
conclusionÀº premises·ÎºÎÅÍ µû¶ó³ª¿Í¾ß ÇÑ´Ù´Â ³í¸®Ãß·Ð |
Induction |
specific case ·ÎºÎÅÍ generalÀ» À¯µµÇس»´Â Ãß·Ð |
Abduction |
true conclusion¿¡¼ ½ÃÀÛÇÏ¿© ±×conclusion ÀÇ ¿øÀÎÀÌ µÇ´Â premises ±îÁö À̸£´Â ¿ª Ãß·Ð. Deduction °ú ¹Ý´ëÀÌ´Ù |
Intuition |
Áõ¸íµÈ ÀÌ·ÐÀÌ ¾ø´Â »óÅÂ. ÀáÀçµÈ ÆÐÅÏÀ» ÀνÄÇØ ³»´Âµ¥ ¹«ÀǽĿ¡ ÀÇÇØ¼¸¸ ´äÀÌ ³ª¿Â´Ù.expert system ¿¡¼´Â ¾ÆÁ÷ ±¸ÇöÇÏÁö ¸øÇÑ´Ù. ¾î¶² Á¶°ÇÀ̳ª ³»»ð¹ý (interpolation) ¿¡ ÀÇÇØ Ã߷еǴ °ÍÀÌ ¾Æ´Ï°í ÈÆ·Ã¿¡ ÀÇÇÑ Ãß·Ð (extrapolate from training)À» Çϱ⠶§¹®¿¡ ANS (ÀÚÀ²½Å°æ°è)¿¡¼³ª °¡´ÉÇÏ´Ù. neural net¿¡¼ ÇØ°á(best guess for a solution) ÇÒ¼ö ÀÖÀ» °ÍÀÌ´Ù. extrapolate : ¿Ü»ð¹ý¿¡ ÀÇÇÑ Ãß·Ð. interpolate : º¸°£¹ý or ³»»ð¹ý¿¡ÀÇÇÑ Ãß·Ð |
Heuristics |
°æÇè¿¡ ¹ÙÅÁÀ» µÐ Rules of thumb |
Generate and Test |
Trial and error¸¦ ÀǹÌ. °¡²û È¿À²¼ºÀ» °Á¶ÇÏ´Â planning¿¡¼ »ç¿ëµÈ´Ù |
Default |
Ưº°ÇÑ Áö½ÄÀÌ ¾ø¾î¼ general or default common knowledge ·Î¼ ÃßÃø |
Autoepistemic |
self-knowledge |
Nonmonotonic |
»õ·Î¿î Áõ°Å¿¡ ÀÇÇØ ±âÁ¸ Áö½ÄÀÌ incorrect ÇØÁú¼ö ÀÖ´Ù |
Analogy |
´Ù¸¥ »óȲ¿¡¼ similarity ¿¡ ±âÃÊÇÑ conclusionÀ» Ãß·ÐÇØ³¿ |
Forward
Chaining °ú Backward Chaining ÀÇ Â÷ÀÌÁ¡
Forward Chaining |
Backward Chaining |
planning, monitoring, control |
diagnosis |
present to future |
present to past |
antecedent to consequent |
consequent to antecedent |
data driven, bottom-up reasoning |
goal driven, top-down reasoning |
work forward to find what solutions follow from the facts |
work backward to find facts that support the hypothesis |
breath-first search facilitated |
depth-first search facilitated |
antecedent determine search |
consequents determine search |
explanation not facilitated |
explanation facilitated |
an interpreter or inference engine |
a database (facts and rules) |
unification À̶ó ºÒ¸®´Â a form of pattern matching |
ÇϳªÀÇ goalÀ» ¸¸Á·ÇÏ´Â search°¡ ½ÇÆÐÇßÀ» ¶§ ¶Ç´Ù¸¥ subgoalÀ» ¼öÇàÇÏ´Â backtracking mechanism |
(Giarratano 1989)
Fuzzy¿Í Neural netÀÇ
°áÇÕ¿¡ °üÇÑ ¿¬±¸
½Å°æ¸ÁÀ» ÀÌ¿ëÇÑ ÆÛÁö ruleÀÇ ÀÚµ¿ÀûÀÎ »ý¼º¿¡ °üÇÑ ¿¬±¸ : ÆÛÁö ½Ã½ºÅÛ ÀÀ¿ëºÐ¾ßÀÇ ½ÇÁ¦ µ¥ÀÌÅ͸¦ ½Å°æ¸ÁÀÇ ÇнÀ µ¥ÀÌÅÍ·Î »ç¿ëÇÏ¿© ±× ÀÀ¿ëºÐ¾ßÀÇ ÆÛÁö ruleÀ» »ý¼ºÇÑ´Ù |
½Å°æ¸Á µµÀÔÀ» ÅëÇÑ ½Å¼ÓÇÑ Ã߷п¡ °üÇÑ ¿¬±¸ : ½Å°æ¸ÁÀº parallel processing°ú multistage Ãß·ÐÀ» ÇÏ´Â ±¸Á¶ÀÌ´Ù. µû¶ó¼ fuzzy¿ÍÀÇ °áÇÕÀº Ãß·Ð °úÁ¤¿¡ ¸¹Àº Á¤º¸°¡ ÇÊ¿äÇÑ °æ¿ì Ãß·Ð ¼Óµµ¸¦ ¸¹ÀÌ °³¼±ÇÒ¼ö ÀÖ´Ù |
ÆÛÁö rule¿¡ Ã߷Рȯ°æ¿¡ ÀûÀÀÇÏ´Â ´É·ÂÀ» ºÎ¿©ÇÏ´Â °Í¿¡ °üÇÑ ¿¬±¸ : Ã߷Рȯ°æÀÌ º¯ÇÔ¿¡ µû¶ó Ãß·Ð rule¿¡ ¹Ý¿µµÉ¼ö ÀÖ´Â »õ·Î¿î ÇнÀ¹æ¹ýÀÇ °³¹ßÀÌ ÇÊ¿äÇÏ´Ù. ÀÌ·³À¸·Î½á »ç¿ëÇÒ¼ö·Ï ¼º´ÉÀÌ ÁÁ¾ÆÁö´Â ½Ã½ºÅÛÀÇ ±¸ÇöÀÌ °¡´ÉÇØ Áø´Ù. À̸¦ À§ÇØ »õ·Î¿î »ç½ÇÀ» ¼ö¿ëÇϰí, º¯ÈµÈ ȯ°æ¿¡ ºÎÀûÀýÇÑ »ç½ÇÀº ¹«½ÃÇØ ¹ö¸®´Â ÇнÀ¹æ¹ýÀÌ ÇÊ¿äÇÏ´Ù.½Å°æ¸Á¿¡¼ ¾ÆÁ÷ À̴ܰè±îÁö´Â °¡Áö ¸øÇß´Ù. |
ÆÛÁö¿¡¼ÀÇ Áö½ÄȹµæÀÇ ¹®Á¦¸¦ ÇØ°á : ½Ã½ºÅÛÀÇ º¹Àâµµ°¡ Áõ°¡ ÇÒ¼ö·Ï Á¶ÀÛÀÚ°¡ ÀÚ½ÅÀÌ ÇàÇÏ´Â Á¦¾î ÇൿÀ» ¾ð¾îÀûÀ¸·Î Ç¥ÇöÇϴµ¥ ÇѰ踦 ´À³¤´Ù. ÀÌÀÇ ÇØ°áÀ» À§ÇØ Á¶ÀÛÀÚÀÇ Çൿ¿¡ µû¸¥ ÀÔÃâ·Â ÀڷḦ ±Ù°Å·Î Á¶ÀýÇàÀ§ÀÇ ÆÛÁö¸ðµ¨À» ¸¸µé°í ½Å°æ¸ÁÀÇ ÇнÀ´É·Â°ú °áÇÕ½ÃŲ´Ù. ÆÛÁöÀ̷аú ½Å°æ¸Á ÀÌ·ÐÀº ¸ðµÎ ƯÁ¤ ºÐ¾ß¿¡ ´ëÇÏ¿© Àΰ£Ã³·³ ÀÏÀ» ÇÒ¼ö ÀÖ´Â ½Ã½ºÅÛÀ» ¸¸µå´Â °Í¿¡ °ü½ÉÀÌ ÀÖ´Ù. ÀÌ ÀÌ·ÐÀ» ÇÔ²² »ç¿ëÇÏ¿© º¸´Ù ³ªÀº ½Ã½ºÅÛÀ» ¸¸µé·Á´Â ¿¬±¸°¡ À§¿Í°°Àº ¹æÇâÀ¸·Î ÀÌ·ç¾î Áö°í ÀÖ´Ù. |
(ÀÌ»ó·É 1995)
ù° : Bayes¸¦ »ç¿ëÇÏ¿© ºÒÈ®½Ç¼ºÀ» 󸮽à ¸ðµç °¡´ÉÇÑ °á°ú°¡ mutually exclusive ÇØ¾ßÇÑ´Ù. ±×·¯³ª ÀÇ·áÁø´ÜÀÇ ½Ç¼¼°è ¹®Á¦¿¡¼´Â ÇѸíÀÇ È¯ÀÚ°¡ µÎ°¡Áö ÀÌ»óÀÇ Áúº´¿¡ °¨¿°µÉ¼ö Àֱ⠶§¹®¿¡ »óÈ£¹èŸÀûÀÌ µÉ¼ö ¾ø´Ù. |
µÑ° : Bayes¿¡¼´Â ÀÌ¹Ì ¾Ë·ÁÁø °¡¼³Áß¿¡¼ Çϳª´Â ¹Ýµå½Ã ÂüÀ̾î¾ß Á¤È®¼ºÀÌ Á¦°íµÈ´Ù. Áï »çÀüÈ®·üÀÌ ÁÖ¾îÁö·Á¸é ÀÌ¹Ì ±× °¡¼³¿¡ ´ëÇÑ Á¤º¸°¡ ÀÖ¾î¾ß ÇÑ´Ù. ±×·¯³ª ¸¸¾à ¾î¶² ȯÀÚ°¡ Áö±Ý±îÁö ¾Ë·ÁÁöÁö ¾ÊÀº Áúº´À» °¡Áö°í ÀÖ¾úÀ» ½Ã¿¡´Â ÁÖ¾îÁø °¡¼³ ¸ðµÎ°¡ °ÅÁþÀϼö ÀÖ´Ù. |
¼Â° : Bayes ¿¡¼´Â Á¶°ÇºÎ È®·üÀ» ÀÌ¿ëÇϹǷΠ»çÀüÈ®·üÀ» ¸ðµÎ ¾Ë°í ÀÖ¾î¾ß Çϳª, À̰ÍÀº Çö½ÇÀûÀ¸·Î ¾î·Á¿ì¸ç ¸¹Àº ¾çÀÇ µ¥ÀÌÅÍ ¹× °æºñ°¡ ¼Ò¿äµÈ´Ù. |
³Ý° : Bayes´Â È®·üÀ» »ç¿ëÇϹǷΠÀüüÀûÀÎ È®·üÀÇ °ªÀº 1 À̵Ǿî¾ß ÇÑ´Ù. ±×·¯¹Ç·Î ¸¸¾à¿¡ »õ·Î¿î Áúº´¿¡ ´ëÇÑ Áö½ÄÀ» database¿¡ »ðÀÔ½Ãų ¶§ ±âÁ¸ÀÇ database¿¡ Á¸ÀçÇÏ´Â È®·ü°ú »õ·Ó°Ô »ðÀÔÇÑ È®·üÀÇ ÇÕÀÌ 1 À̵ǵµ·Ï °¢°¢ÀÇ È®·üÀ» ¼öÁ¤ÇØ¾ß ÇÑ´Ù. ÀÌ·¯ÇÑ ¼öÁ¤¹æ¹ýÀº ¸Å¿ì ¾î·Á¿ì¸ç º¹ÀâÇÑ °úÁ¤À» °ÅÄ£´Ù |
(Shortliffe ±èȼö 1993)