• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) Áö½Ä ÃßÃâÀ» À§ÇÑ È¿À²ÀûÀÎ ´ÙÁß ÀÛ¾÷ Å©¶ó¿ìµå¼Ò½Ì ±â¹ý
¿µ¹®Á¦¸ñ(English Title) Cost-effective Multi-task Crowdsourcing Method for Knowledge Extraction
ÀúÀÚ(Author) ³²»óÇÏ   À̹ÎÈ£   ÇãöÈÆ   Ãֱ⼱   Sangha Nam   Minho Lee   Cheolhoon Heo   Key-Sun Choi  
¿ø¹®¼ö·Ïó(Citation) VOL 26 NO. 11 PP. 0507 ~ 0512 (2020. 11)
Çѱ۳»¿ë
(Korean Abstract)
Áö½Ä ÃßÃâÀ̶õ ÀÚ¿¬¾ð¾î ¹®ÀåÀ¸·ÎºÎÅÍ ÄÄÇ»ÅÍ°¡ ÀÌÇØÇÒ ¼ö ÀÖ´Â Áö½Ä ÇüÅÂÀÇ Á¤º¸¸¦ ÃßÃâÇÏ´Â ÀÛ¾÷À¸·Î, Áö½Äº£À̽º¸¦ »ý¼º ¹× È®ÀåÇÏ°í À̸¦ ÀÌ¿ëÇØ ÁúÀÇ ÀÀ´ä, ´ëÈ­ ¿¡ÀÌÀüÆ® µî ´Ù¾çÇÑ ÀÀ¿ë ºÐ¾ß¿¡ È°¿ëµÈ´Ù. Áö½Ä ÃßÃâÀ» À§Çؼ­´Â ÁÖ¾îÁø ÀÚ¿¬¾ð¾î ¹®Àå ³»¿¡ Áö½Äº£À̽º¿¡ ¿¬°á °¡´ÉÇÑ °³Ã¼¸¦ ¹ß°ßÇÏ°í ±× °³Ã¼µé °£ÀÇ ÀǹÌÀû °ü°è¸¦ ÆľÇÇÏ´Â °úÁ¤ÀÌ ÇʼöÀûÀ¸·Î ¼ö¹ÝµÈ´Ù. º» ³í¹®¿¡¼­´Â Å©¶ó¿ìµå¼Ò½ÌÀ¸·Î °³Ã¼ ¿¬°á°ú °ü°è ÃßÃâÀ» À§ÇÑ µ¥ÀÌÅÍ ¼ÂÀ» È¿°úÀûÀ¸·Î Á¦ÀÛÇÏ´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. Áö½Ä ÃßÃâ¿¡´Â ´ÙÁß ÀÛ¾÷ ¼³°è°¡ ÇÊ¿äÇѵ¥ È¿À²ÀûÀÎ µ¥ÀÌÅÍ ÁÖ¼® ÀÛ¾÷ ƲÀ» ¼³°èÇÏ°í, ¶ÇÇÑ ¾çÁúÀÇ µ¥ÀÌÅ͸¦ ¼öÁýÇϱâ À§ÇØ ÀÛ¾÷ÀÚµéÀ» ±³À°, µ¥ÀÌÅÍ Ç°ÁúÀ» Áö¼ÓÀûÀ¸·Î ÀÚµ¿ °Ë¼ö, ±×¸®°í ¾Ç¼º ÀÛ¾÷ÀÚ¸¦ °É·¯³»´Â ÀåÄ¡µéÀ» µµÀÔÇÏ¿´´Ù. º» ³í¹®¿¡¼­ ¼öÁýÇÑ µ¥ÀÌÅÍ¿Í ¹æ¹ýÀÇ ¿ì¼ö¼ºÀ» ÀÔÁõÇϱâ À§ÇØ, Áö½Ä ÃßÃâ¿¡ ÇÊ¿äÇÑ ¿©·¯ °¡Áö ¸ðµ¨µéÀ» ÇнÀÇÏ°í Æò°¡ÇÑ °á°ú, º» ³í¹®ÀÇ Å©¶ó¿ìµå¼Ò½Ì µ¥ÀÌÅÍ°¡ ¼º´É Çâ»ó¿¡ ±â¿©Çß´Ù´Â °ÍÀ» È®ÀÎÇÏ¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Knowledge extraction is the task of extracting knowledge from natural language sentences in the form of information that a computer can understand to create and expand the knowledge base. Knowledge bases and extraction systems are used in various applications such as question answering and dialog agents. For knowledge extraction, it is essential to both find a knowledge base entity in a given natural language sentence and understand the semantic relationship between the entities. In this paper, we proposed a cost-effective multi-task data collection process to create a dataset for entity linking, entity grouping, and relation extraction using crowdsourcing. Because knowledge extraction requires multiple tasks, we designed an annotation scheme to reduce the burden of annotation by humans in each step where the output of the previous task was used as the input for the next task. We trained, tested, and continuously monitored the workers to filter out malicious workers. To demonstrate the excellence of our data and methods, the results of learning and evaluating the models for knowledge extraction confirmed that our crowdsourcing scheme contributed to performance improvement.
Å°¿öµå(Keyword) ½Ã¸ÇƽÀ¥   Áö½Äº£À̽º   ÀÚ¿¬¾ð¾î󸮠  Áö½ÄÃßÃâ   Å©¶ó¿ìµå¼Ò½Ì   semantic web   knowledge base   natural language processing   knowledge extraction crowdsourcing  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå