ÀÚ·á¼³¸í
1. 1. Ž»ö ¾Ë°í¸®Áò¿¡ ´ëÇÑ ´ÙÀ½ Áú¹®¿¡ ´äÇϽÿÀ. (20)
(1) ÀÇ evaluation function, ÀÇ ÇÔ¼ö½ÄÀ» ¾²°í, ÀÌ¿¡ ...
º»¹®/³»¿ë
1. 1. Ž»ö ¾Ë°í¸®Áò¿¡ ´ëÇÑ ´ÙÀ½ Áú¹®¿¡ ´äÇϽÿÀ. (20)
(1) ÀÇ evaluation function, ÀÇ ÇÔ¼ö½ÄÀ» ¾²°í, ÀÌ¿¡ ´ëÇØ ¼³¸íÇϽÿÀ.
(2) ÀÇ °¡Àå Å« Ư¡Àº admissibilityÀÌ´Ù. AdmissibilityÀÇ ¶æÀº?
[Ç®ÀÌ]
(1)
: Áö±Ý±îÁö ¹ß°ßµÈ start node·ÎºÎÅÍ node ±îÁöÀÇ Ãִܰæ·ÎÀÇ ºñ¿ë
: node À¸·ÎºÎÅÍ goal node±îÁöÀÇ expected cost. ½ÇÁ¦ ºñ¿ëÀ» À̶ó°í ÇÒ ¶§, À» ¸¸Á·ÇØ¾ß ÇÑ´Ù.
(2) ÇØ°¡ ÀÖÀ¸¸é ¹Ýµå½Ã ÇØ¸¦ ãÀ¸¸ç, ãÀº ÇØ´Â optimalÇÏ´Ù.
2. Consider the following game tree in which the static scores are all from the first player`s point of view. Assume that the first player is the MAX. (20)
7 2 1 8 9 1 8 3 3 8
1. k-nearest neighbor learning¿¡ ´ëÇØ ¼³¸íÇϽÿÀ. (10)
[Ç®ÀÌ]
1. ÇнÀ½Ã¿¡´Â ÈÆ·Ã µ¥ÀÌÅ͵éÀ» ±×´ë·Î ¿Ü¿î´Ù.
2. »õ·Î¿î µ¥ÀÌÅÍ¿¡ ´ëÇØ¼ ÆÇ´ÜÀ» ÇÒ ¶§¿¡´Â, ±â¾ïµÈ µ¥ÀÌÅÍ Áß »õ·Î¿î µ¥ÀÌÅÍ¿Í °¡Àå °¡±î¿î °³ÀÇ µ¥ÀÌÅ͸¦ ã¾Æ¼ ÀÌ °³ÀÇ µ¥ÀÌÅ͵éÀÇ Å¬·¡½ºµé Áß °¡Àå ¿ì¼¼ÇÑ Å¬·¡½º¸¦ ¡¦(»ý·«)
3. ½Å°æÈ¸·Î¸ÁÀÇ generalizationÀ» ³ôÀ̱â À§Çؼ´Â hidden unitÀÇ °³¼ö¸¦ º¯È½ÃÄѰ¡¸é¼ °¢ °æ¿ì¸¶´Ù generalization error¸¦ ÃøÁ¤ÇÏ¿© ÀÌ °ªÀÌ ÃÖ¼Ò°¡ µÇ´Â hidden unitÀÇ °³¼ö¸¦ ¼±ÅÃÇÏ´Â ¹æ¹ýÀ» »ç¿ëÇÒ ¼ö ÀÖ´Ù. ´ÙÀ½ ¹°À½¿¡ ´äÇϽÿÀ.(20)
(1) ±â°è ÇнÀ¿¡ ÀÖ¾î¼ generalizationÀ̶õ ¹«¾ùÀΰ¡?