Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö ¼ÒÇÁÆ®¿þ¾î ¹× µ¥ÀÌÅÍ °øÇÐ
ÇѱÛÁ¦¸ñ(Korean Title) |
Sequence-to-Sequence ModelÀ» ÀÌ¿ëÇÑ ¿µ¾î ¹ßÀ½ ±âÈ£ ÀÚµ¿ º¯È¯ |
¿µ¹®Á¦¸ñ(English Title) |
Automatic Conversion of English Pronunciation Using Sequence-to-Sequence Model |
ÀúÀÚ(Author) |
ÀÌ°øÁÖ
ÃÖ¿ë¼®
Kong Joo Lee
Yong Seok Choi
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 06 NO. 05 PP. 0267 ~ 0278 (2017. 05) |
Çѱ۳»¿ë (Korean Abstract) |
¿µ¾î´Â µ¿ÀÏ Ã¶ÀÚÀÇ ¹ßÀ½ÀÌ ¸Å¿ì ´Ù¾çÇÑ ¾ð¾îÀ̱⠶§¹®¿¡ »çÀü¿¡ ±â¼úµÇ¾î ÀÖ´Â ´Ü¾îÀÇ ¹ßÀ½±âÈ£¸¦ Àоî¾ß¸¸ Á¤È®ÇÑ ¹ßÀ½À» ¾Ë ¼ö ÀÖ´Ù. ¿µ¾î »çÀü¸¶´Ù »ç¿ëÇÏ´Â ¹ßÀ½±âÈ£(phonetic alphabet) ½Ã½ºÅÛÀÌ ´Ù¸£¸ç °°Àº ´Ü¾î¿¡ ´ëÇØ ±â¼úÇÏ°í ÀÖ´Â ¹ßÀ½ ¿ª½Ã ´Ù¸£´Ù. º» ¿¬±¸¿¡¼´Â ÃÖ±Ù µö ·¯´× ºÐ¾ß¿¡¼ ³Î¸® »ç¿ëµÇ°í ÀÖ´Â sequence-to-sequence (seq2seq) modelÀ» ÀÌ¿ëÇÏ¿© »çÀü¸¶´Ù ´Ù¸¥ ¹ßÀ½À» ÀÚµ¿À¸·Î º¯È¯ÇØ º¸°íÀÚ ÇÑ´Ù. 4°¡Áö ´Ù¸¥ Á¾·ùÀÇ »çÀü¿¡¼ ÃßÃâÇÑ ¹ßÀ½ µ¥ÀÌÅ͸¦ ÀÌ¿ëÇÏ¿© ¸ðµÎ 12°³ÀÇ seq2seq modelÀ» ±¸ÇöÇÏ¿´À¸¸ç, ¹ßÀ½ ÀÚµ¿ º¯È¯ ¸ðµâÀÇ Á¤È® ÀÏÄ¡À² Àº 74.5% ¢¦ 89.6%ÀÇ ¼º´ÉÀ» º¸¿´´Ù. º» ¿¬±¸ÀÇ ÁÖ¿ä ¸ñÀûÀº ´ÙÀ½ÀÇ µÎ °¡ÁöÀÌ´Ù. ù° ¿µ¾î ¹ßÀ½±âÈ£ ½Ã½ºÅÛ°ú °¢ »çÀüÀÇ ¹ßÀ½ µ¥ÀÌÅÍ Æ¯¼ºÀ» »ìÆ캸´Â °ÍÀÌ°í, µÑ°, ¹ßÀ½ Á¤º¸ÀÇ ÀÚµ¿ º¯È¯°ú ¿À·ù ºÐ¼®À» ÅëÇØ seq2seq modelÀÇ Æ¯¼ºÀ» »ìÆ캸´Â °ÍÀÌ´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
As the same letter can be pronounced differently depending on word contexts, one should refer to a lexicon in order to pronounce a word correctly. Phonetic alphabets that lexicons adopt as well as pronunciations that lexicons describe for the same word can be different from lexicon to lexicon. In this paper, we use a sequence-to-sequence model that is widely used in deep learning research area in order to convert automatically from one pronunciation to another. The 12 seq2seq models are implemented based on pronunciation training data collected from 4 different lexicons. The exact accuracy of the models ranges from 74.5% to 89.6%. The aim of this study is the following two things. One is to comprehend a property of phonetic alphabets and pronunciations used in various lexicons. The other is to understand characteristics of seq2seq models by analyzing an error.
|
Å°¿öµå(Keyword) |
¹ßÀ½±âÈ£
´Ü¾î ¹ßÀ½
¿µ¾î ´Ü¾î ¹ßÀ½ÀÇ ´Ù¾ç¼º
Sequence-to-Sequence Model
Phonetic Alphabet
Word¡¯s Pronunciation
Allophone of English Word
Sequence-to-Sequence Model
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|