• 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) STDP ¾Ë°í¸®Áò°ú ½ºÆÄÀÌÅ© °£ÀÇ ½Ã°£Àû »óÈ£ ÀÛ¿ë¿¡ µû¸¥ SNNÀÇ ÇнÀ ¼º´É ¹× ½Ã°£ ºÐ¼®
¿µ¹®Á¦¸ñ(English Title) Analysis of the Training Performance and Time of SNN by STDP Algorithms and Spike Temporal Interactions
ÀúÀÚ(Author) ¹Ú¼º½Ä   À±¼º·Î   Seongsik Park   Sungroh Yoon  
¿ø¹®¼ö·Ïó(Citation) VOL 24 NO. 09 PP. 0482 ~ 0486 (2018. 09)
Çѱ۳»¿ë
(Korean Abstract)
¿©·¯ ºÐ¾ß¿¡¼­ Àΰø ½Å°æ¸ÁÀ» »ç¿ëÇÑ µö ·¯´×ÀÌ ±âÁ¸ ´Ù¸¥ ¾Ë°í¸®ÁòÀÇ ¼º´ÉÀ» Å©°Ô ¶Ù¾î ³ÑÀ¸¸é¼­ Å« °ü½ÉÀ» ¹Þ°í ÀÖ´Ù. ÇÏÁö¸¸, ÇöÀç ÁÖ·Î »ç¿ëµÇ´Â µö ·¯´× ¹æ½ÄÀº Àü·Â ¼Ò¸ð ¿ä±¸·®ÀÌ Å©±â ¶§¹®¿¡ Á¦ÇÑÀûÀÎ ÀÚ¿øÀ» °®°í ÀÖ´Â ¸ð¹ÙÀÏ ºÐ¾ß¿¡ Àû¿ëµÇ±â ¾î·Æ´Ù. ÀÌ¿¡ µû¶ó Àú Àü·ÂÀ¸·Î µ¿ÀÛÇÒ ¼ö ÀÖ´Â spiking neural networks (SNNs)¿¡ ´ëÇÑ °ü½ÉÀÌ Ä¿Áö°í ÀÖ´Ù. SNNÀº ½Ã³À½º Àü°ú ÈÄÀÇ ½ºÆÄÀÌÅ© ½Ã°£°ü°è¿¡ µû¶ó ½Ã³À½º °¡ÁßÄ¡°¡ Á¶ÀýµÇ´Â STDP ¾Ë°í¸®ÁòÀ» »ç¿ëÇÏ¿© ½Ã³À½º °¡ÁßÄ¡¸¦ ÇнÀÇÑ´Ù. µû¶ó¼­ SNNÀº ÇнÀ¿¡ »ç¿ëÇÏ´Â ½ºÆÄÀÌÅ©ÀÇ ¼ö¿¡ µû¸¥ STDP ¾Ë°í¸®Áò°ú ½ºÆÄÀÌÅ© °£ÀÇ ½Ã°£Àû »óÈ£ ÀÛ¿ë¿¡ µû¶ó ´Ù¾çÇÑ ±¸¼ºÀ¸·Î ÇнÀ ÇÒ ¼ö ÀÖ´Ù. º» ³í¹®¿¡¼­´Â ¿©·¯ STDP ¾Ë°í¸®Áò ±¸¼ºÀ¸·Î SNNÀ» ÇнÀÇÏ°í ÇнÀ ¼º´É°ú ÇнÀ ½Ã°£À» ºñ±³ÇØ º¸¾Ò´Ù.
¿µ¹®³»¿ë
(English Abstract)
Deep learning using artificial neural networks in various fields has received great attention because it outperforms other existing algorithms. However, deep learning cannot be applied to a mobile environment with limited hardware resources mainly because of power issues. Thus, there has been growing interest in spiking neural networks (SNNs) that can operate with low power consumption. SNNs learn synaptic weights by STDP algorithms, which adjust the synaptic weights according to the time dependency of the pre- and post-spikes. We can train SNNs in various configurations of STDP algorithms and spike temporal interactions. In this paper, we trained an SNN with the configurations of training algorithms and evaluated the performance and time of training the SNN.
Å°¿öµå(Keyword) SNN   STDP   artificial intelligence   spike temporal interaction  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå