Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
Èñ¼Ò µ¥ÀÌÅ͸¦ À§ÇÑ °ÀÎ ¼Õ½Ç ÇÔ¼ö¸¦ ÀÌ¿ëÇÑ ÁØ Áöµµ ÇнÀ |
¿µ¹®Á¦¸ñ(English Title) |
Semi-Supervised Learning Exploiting Robust Loss Function for Sparse Labeled Data |
ÀúÀÚ(Author) |
ÃÖ½½±â
±è¿µÇ¥
Ȳ¼ÒÁ¤
ÃÖÈñ¿
Seulgi Choi
Youngpyo Kim
Sojung Hwang
Heeyoul Choi
¾È¿µÁØ
½É±Ô¼®
Youngjun Ahn
Kyuseok Shim
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 48 NO. 12 PP. 1343 ~ 1348 (2021. 12) |
Çѱ۳»¿ë (Korean Abstract) |
ÀÌ ³í¹®¿¡¼´Â µ¥ÀÌÅÍÀÇ ·¹À̺íÀÌ ¸Å¿ì ºÎÁ·ÇÑ »óȲ¿¡¼ µ¥ÀÌÅÍ Áõ°±â¹ý°ú °ÀÎ ¼Õ½Ç ÇÔ¼ö¸¦ »ç¿ëÇÏ¿© ÁØ Áöµµ ÇнÀÀ» ÇÏ´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ±âÁ¸ µ¥ÀÌÅÍ Áõ°±â¹ýÀ» »ç¿ëÇÏ´Â ÁØ Áöµµ ÇнÀ ¹æ¹ýÀº ·¹À̺íÀÌ ¾ø´Â µ¥ÀÌÅ͸¦ Áõ°ÇÏ°í, ±× Áß ½Å·Úµµ°¡ ³ôÀº µ¥ÀÌÅÍ¿¡ ´ëÇؼ¸¸ ÇöÀç ¸ðµ¨ÀÌ ¿¹ÃøÇÑ ·¹À̺íÀ» ¿ø ÇÖ º¤ÅÍ·Î ºÙ¿© ÇнÀ¿¡ »ç¿ëÇÑ´Ù. ±×·¡¼ ½Å·Úµµ°¡ ³·Àº µ¥ÀÌÅÍ´Â »ç¿ëÇÏÁö ¾Ê´Â ¹®Á¦°¡ ÀÖ¾ú´Âµ¥, À̸¦ ÇØ°áÇϱâ À§ÇØ °ÀÎ ¼Õ½Ç ÇÔ¼ö¸¦ ÀÌ¿ëÇÏ¿© ½Å·Úµµ°¡ ³·Àº µ¥ÀÌÅÍ ¶ÇÇÑ »ç¿ëÇÏ´Â ¿¬±¸µµ ÁøÇàµÇ¾ú´Ù. ÇÑÆí, ·¹À̺íÀÌ ¸Å¿ì ÀûÀº »óȲ¿¡¼´Â ¸ðµ¨ÀÌ ¿¹ÃøÇÑ ·¹À̺íÀº ½Å·Úµµ°¡ ³ô´õ¶óµµ ºÎÁ¤È®ÇÏ´Ù´Â ¹®Á¦°¡ ÀÖ´Ù. ÀÌ ³í¹®¿¡¼´Â ·¹À̺íÀÌ ¸Å¿ì ÀûÀº »óȲ¿¡¼ ¿ø ÇÖ º¤ÅÍ°¡ ¾Æ´Ñ ¸ðµ¨ÀÌ ¿¹ÃøÇÑ È®·üÀ» ·¹À̺í·Î »ç¿ëÇÔÀ¸·Î½á ºÐ·ù ¸ðµ¨ÀÇ ¼º´ÉÀ» ³ôÀÏ ¼ö ÀÖ´Â ¹æ¹ýÀ» Á¦½ÃÇÑ´Ù. ¶ÇÇÑ À̹ÌÁö ºÐ·ù ¹®Á¦¿¡ ´ëÇÑ ½ÇÇèÀ» ÅëÇÏ¿© Á¦½ÃµÈ ¹æ¹ýÀÌ ºÐ·ù ¸ðµ¨ÀÇ ¼º´ÉÀ» Çâ»ó½ÃÅ´À» º¸¿©ÁØ´Ù. |
¿µ¹®³»¿ë (English Abstract) |
This paper proposes a semi-supervised learning method which uses data augmentation and robust loss function when labeled data are extremely sparse. Existing semi-supervised learning methods augment unlabeled data and use one-hot vector labels predicted by the current model if the confidence of the prediction is high. Since it does not use low-confidence data, a recent work has used low-confidence data in the training by utilizing robust loss function. Meanwhile, if labeled data are extremely sparse, the prediction can be incorrect even if the confidence is high. In this paper, we propose a method to improve the performance of a classification model when labeled data are extremely sparse by using predicted probability, instead of one hot vector as the label. Experiments show that the proposed method improves the performance of a classification model. |
Å°¿öµå(Keyword) |
±³À°
¸ÂÃãÇü ÇнÀ
Áö½Ä ÃßÀû
µö·¯´×
ÀÚ±â ÁÖÀÇ ÁýÁß
education
personalized learning
knowledge tracing
deep learning
self-attention
µö ·¯´×
ÁØ Áöµµ ÇнÀ
µ¥ÀÌÅÍ Áõ°
°ÀÎ ¼Õ½Ç ÇÔ¼ö
deep learning
semi-supervised learning
data augmentation
robust loss function
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|