Ç¥ÁØÈ­ Âü¿©¾È³»

TTAÀÇ Ç¥ÁØÇöȲ

Ȩ > Ç¥ÁØÈ­ °³¿ä > TTAÀÇ Ç¥ÁØÇöȲ

Ç¥ÁعøÈ£ TTAK.KO-10.1419 ±¸Ç¥ÁعøÈ£
Á¦°³Á¤ÀÏ 2023-06-30 ÃÑÆäÀÌÁö 43
ÇѱÛÇ¥Áظí Çѱ¹¾î À½¼º ¹× ÅؽºÆ® µ¥ÀÌÅÍÀÇ ÀǹÌÀû Á¤È®¼º Ç°Áú°ËÁõ ¹æ¹ý
¿µ¹®Ç¥Áظí Quality Verification for Semantic Accuracy of Voice and Text Data in Korean
Çѱ۳»¿ë¿ä¾à ÀÌ Ç¥ÁØÀº ÀΰøÁö´É ÇнÀÀ» À§ÇÑ À½¼º ¹× ÅؽºÆ® µ¥ÀÌÅÍÀÇ Ç°Áú°ËÁõ ¹æ½Ä¿¡ ´ëÇÑ º¸ÆíÀû °üÁ¡ÀÇ Ã¼°èÈ­µÈ Ç°Áú°ËÁõ ±âÁØ°ú ¹æ½ÄÀ» Á¦½ÃÇÑ´Ù. ƯÈ÷ ÀΰøÁö´É È°¿ë¼­ºñ½º °üÁ¡ÀÇ ÀÓ¹« ´Þ¼ºÀ» À§ÇÏ¿© µ¥ÀÌÅÍ ¼³°è Àǵµ¿¡ ¸Â°Ô ¶óº§¸µ ÀÛ¾÷ÀÌ Àß ¼öÇàµÇ¾ú´ÂÁö¿¡ ´ëÇÑ ÀǹÌÀû Á¤È®¼ºÀ» Æò°¡ÇÑ´Ù. À½¼º µ¥ÀÌÅÍ´Â À½¼º ÀÎ½Ä ÀÓ¹«¸¦ À§ÇÑ ¶óº§¸µÀÇ ÀǹÌÀû Á¤È®¼ºÀ» Æò°¡ÇÏ°í, ÅؽºÆ® µ¥ÀÌÅÍ´Â ÁúÀÇÀÀ´ä, ¿ä¾à, ¹ø¿ª, ´ëÈ­, ¸»¹¶Ä¡ ÀÓ¹«¸¦ À§ÇÑ ¶óº§¸µÀÇ ÀǹÌÀû Á¤È®¼ºÀ» Æò°¡ÇÏ´Â º¸ÆíÀûÀÎ Ç°Áú°ËÁõ ¹æ¹ýÀ» ±â¼úÇÑ´Ù.
ÀÓ¹« ÀÇÁ¸Àû µ¥ÀÌÅÍ Ç°ÁúÀº µ¥ÀÌÅ͸¦ È°¿ëÇÏ°íÀÚ ÇÏ´Â ÀÓ¹«º° °íÀ¯ÀÇ Ç°Áú ±âÁØÀ» ÁؼöÇØ¾ß ÇÔÀ» ÀǹÌÇÑ´Ù. ÀÌ Ç¥ÁØÀº À½¼º ¹× ÅؽºÆ® µ¥ÀÌÅÍÀÇ Ç°Áú°ËÁõ °üÁ¡À» ÅëÇØ º¸ÆíÀûÀÌ°í È°¿ë ¸ñÀû¿¡ ÀûÇÕÇÑ Ç°Áú ¿ä¼Ò¸¦ Á¦½ÃÇÏ°í, Ç°Áú°ËÁõ ±âÁØ°ú ¹æ¹ýÀ» Á¦°øÇÑ´Ù.
¿µ¹®³»¿ë¿ä¾à This standard presents standardized quality verification criteria and methods from a universal perspective for the quality verification methods of voice and text data for artificial intelligence learning. In particular, it evaluates the semantic accuracy of labeling tasks performed in accordance with data design intentions, for achieving missions from the perspective of AI utilization services. Voice data evaluates the semantic accuracy of labeling for speech recognition tasks, and text data describes universal quality verification methods for evaluating the semantic accuracy of labeling for tasks such as question-answering, summarization, translation, conversation, and corpus tasks.
±¹Á¦Ç¥ÁØ
°ü·ÃÆÄÀÏ TTAK_[1].KO-10.1419.pdf TTAK_[1].KO-10.1419.pdf            

ÀÌÀü
¿øÀÚ·Â ¹ßÀü ¼³ºñ °ü¸®¸¦ À§ÇÑ »ç¹°ÀÎÅÍ³Ý ¹«¼± ³×Æ®¿öÅ© - Á¦2ºÎ: À̵¿Çü ¹æ»ç¼±·®ÀåÄ¡ ÀÎÅÍÆäÀ̽º
´ÙÀ½
ÄÄÇ»ÅÍ ºñÀü ºÐ¾ß ¶óº§µ¥ÀÌÅÍÀÇ ÀǹÌÀû Á¤È®¼º ¹× À¯È¿¼º Ç°Áú°ËÁõ ¹æ¹ý