Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)
ÇѱÛÁ¦¸ñ(Korean Title) |
½Ã°è¿ ½ÉÃþÇнÀ ¸ðµ¨ÀÇ Àº´Ð ³ëµå¿¡ ´ëÇÑ ½Ã°¢È |
¿µ¹®Á¦¸ñ(English Title) |
Visualization of Convolutional Neural Networks for Time Series Input Data |
ÀúÀÚ(Author) |
Á¶¼ÒÈñ
ÃÖÀç½Ä
Sohee Cho
Jaesik Choi
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 47 NO. 05 PP. 0445 ~ 0453 (2020. 05) |
Çѱ۳»¿ë (Korean Abstract) |
»ê¾÷, ÀÇ·á, ±ÝÀ¶ µî ´Ù¾çÇÑ ºÐ¾ß¿¡¼ ÀΰøÁö´ÉÀ» È°¿ëÇÑ ¿¹Ãø ¹× Áø´ÜÀÌ ´Ã¾î³ª¸é¼, ÀΰøÁö´ÉÀÇ ³»ºÎ ÀÛµ¿¿ø¸®¸¦ ¼³¸íÇÏ´Â ¿¬±¸¿¡µµ °ü½ÉÀÌ ³ô¾ÆÁö°í ÀÖ´Ù. À̹ÌÁö µ¥ÀÌÅÍ¿¡¼ Áß¿ä ÀԷ Ư¡Á¡À» ½Ã°¢ÈÇÏ´Â ±âÁ¸ ¿¬±¸µé°ú ´Ù¸£°Ô, º» ³í¹®¿¡¼´Â ½Ã°è¿ µ¥ÀÌÅÍÀÇ Àº´Ð ³ëµå¸¦ ½Ã°¢ÈÇÏ¿© ½ÉÃþ½Å°æ¸Á ³»ºÎÀÇ ÀÛµ¿¿ø¸®¸¦ ¼³¸íÇÑ´Ù. º» ³í¹®Àº Àº´Ð ³ëµåÀÇ ½Ã°¢È¸¦ ½±°Ô Çϵµ·Ï °¡ÁßÄ¡ Çà·Ä(weight matrix)À» ±âÁØÀ¸·Î Àº´Ð ³ëµå¸¦ ±ºÁýÈÇÏ¿© ÆÐÅÏÀ» ÆľÇÇÏ¿´´Ù. À̸¦ ÅëÇØ ½ÉÃþÇнÀ ¸ðµ¨ÀÇ ÀÛµ¿¿ø¸®¸¦ ¼³¸íÇÒ »Ó¸¸ ¾Æ´Ï¶ó, »ç¿ëÀÚ ¼öÁØ¿¡¼ ½Ã°è¿ µ¥ÀÌÅÍ¿¡ ´ëÇÑ ÀÌÇظ¦ ³ôÀÏ ¼ö ÀÖ¾ú´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
Globally, the use of artificial intelligence (AI) applications has increased in a variety of industries from manufacturing, to health care to the financial sector. As a result, there is a growing interest in explainable artificial intelligence (XAI), which can provide explanations of what happens inside AI. Unlike previous work using image data, we visualize hidden nodes for a time series. To interpret which patterns of a node make more effective model decisions, we propose a method of arranging nodes in a hidden layer. The hidden nodes sorted by weight matrix values show which patterns significantly affected the classification. Visualizing hidden nodes explains a process inside the deep learning model, as well as enables the users to improve their understanding of time series data.
|
Å°¿öµå(Keyword) |
½Ã°è¿ µ¥ÀÌÅÍ
½ÉÃþÇнÀ
Àº´Ð ³ëµå
time series data
deep learning
hidden layers
weight matrix
patterns
°¡ÁßÄ¡ Çà·Ä
ÆÐÅÏ
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|