• 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¸°úÇÐȸ³í¹®Áö (Journal of KIISE)

Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)

Current Result Document : 29 / 29

ÇѱÛÁ¦¸ñ(Korean Title) µ¥ÀÌÅÍ ½ºÆ®¸² ºÐ·ù¸¦ À§ÇÑ µö·¯´× Ãß·Ð ¸ðµ¨ÀÇ ºÐ»ê ó¸®
¿µ¹®Á¦¸ñ(English Title) Distributed Processing of Deep Learning Inference Models for Data Stream Classification
ÀúÀÚ(Author) ¹®È¿Á¾   ¼Õ½Ã¿î   ¹®¾ç¼¼   Hyojong Moon   Siwoon Son   Yang-Sae Moon                          
¿ø¹®¼ö·Ïó(Citation) VOL 48 NO. 10 PP. 1154 ~ 1165 (2021. 10)
Çѱ۳»¿ë
(Korean Abstract)
´Ù¾çÇÑ ºÐ¾ß¿¡¼­ µ¥ÀÌÅÍ ½ºÆ®¸²ÀÌ »ý¼ºµÇ°í ÀÖÀ¸¸ç, À̸¦ µö·¯´×¿¡ Àû¿ëÇÏ´Â È°¿ë »ç·Ê°¡ Áõ°¡ÇÏ°í ÀÖ´Ù. µö·¯´×À» »ç¿ëÇÏ¿© µ¥ÀÌÅÍ ½ºÆ®¸²À» ºÐ·ùÇϱâ À§Çؼ­´Â ¼­ºù(serving)À» ÅëÇØ ¸ðµ¨À» ½Ç½Ã°£ ½ÇÇà½ÃÄÑ¾ß ÇÑ´Ù. ÀÌ·¯ÇÑ ¼­ºù ¸ðµ¨Àº gRPC ¶Ç´Â HTTP Åë½ÅÀ¸·Î ÀÎÇØ µ¥ÀÌÅÍ ½ºÆ®¸²À» ºÐ·ù¿¡ Å« Áö¿¬ ½Ã°£ÀÌ ¹ß»ýÇÑ´Ù. ¶ÇÇÑ, ¼­ºùµÈ ¸ðµ¨ÀÌ ³ôÀº º¹Àâµµ¸¦ °¡Áö´Â ½ºÅÂÅ· Ãß·Ð ¸ðµ¨À̶ó¸é, µ¥ÀÌÅÍ ½ºÆ®¸² ºÐ·ù¿¡ ´õ Å« Áö¿¬½Ã°£ÀÌ ¹ß»ýÇÑ´Ù. À̸¦ ÇØ°áÇϱâ À§ÇØ, º» ³í¹®¿¡¼­´Â ¾ÆÆÄÄ¡ ½ºÅè(Apache Storm)À» »ç¿ëÇÑ µ¥ÀÌÅÍ ½ºÆ®¸² ºÐ·ùÀÇ ºÐ»ê ó¸® ÇØ°áÃ¥À» Á¦¾ÈÇÑ´Ù. ù°, ±âÁ¸ ¼­ºù ¹æ¹ýÀ¸·Î µ¥ÀÌÅÍ ½ºÆ®¸²À» ºÐ·ùÇÒ ¶§ ¹ß»ýÇÏ´Â Áö¿¬½Ã°£À» ÁÙÀ̱â À§ÇØ ¾ÆÆÄÄ¡ ½ºÅè ±â¹Ý ½Ç½Ã°£ ºÐ»ê Ãß·Ð ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. ½ÇÇè °á°ú, Á¦¾ÈÇÑ ºÐ»ê Ãß·Ð ±â¹ýÀÌ ±âÁ¸ ¼­ºù ¹æ¹ý¿¡ ºñÇØ ÃÖ´ë 11¹è±îÁö Áö¿¬½Ã°£À» ÁÙÀÎ °ÍÀ¸·Î ³ªÅ¸³µ´Ù. µÑ°, ½ºÅÂÅ·À» Àû¿ëÇÑ ¾Ç¼º URL ŽÁö ¸ðµ¨·Î URL ½ºÆ®¸²À» ºÐ·ùÇÒ ¶§ÀÇ Áö¿¬½Ã°£À» ÁÙÀ̱â À§ÇØ, ³× °¡Áö ºÐ»ê ó¸® ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. Á¦¾ÈÇÏ´Â ºÐ»ê ó¸® ±â¹ýÀº Independent Stacking, Sequential Stacking, Semi- Sequential Stacking, Stepwise-Independent StackingÀÌ´Ù. ½ÇÇè °á°ú, µ¶¸³Àû ¼öÇà°ú ¼øÂ÷Àû ó¸®ÀÇ Æ¯¼ºÀ» °¡Áø Stepwise-Independent StackingÀÌ °¡Àå ÀÛÀº Áö¿¬½Ã°£À» º¸¿©, URL ½ºÆ®¸² ºÐ·ù¿¡ °¡Àå ÀûÇÕÇÑ °ÍÀ¸·Î ³ªÅ¸³µ´Ù.
¿µ¹®³»¿ë
(English Abstract)
The increased generation of data streams has subsequently led to increased utilization of deep learning. In order to classify data streams using deep learning, we need to execute the model in real-time through serving. Unfortunately, the serving model incurs long latency due to gRPC or HTTP communication. In addition, if the serving model uses a stacking ensemble method with high complexity, a longer latency occurs. To solve the long latency challenge, we proposed distributed processing solutions for data stream classification using Apache Storm. First, we proposed a real-time distributed inference method based on Apache Storm to reduce the long latency of the existing serving method. The present study's experimental results showed that the proposed distributed inference method reduces the latency by up to 11 times compared to the existing serving method. Second, to reduce the long latency of the stacking-based inference model for detecting malicious URLs, we proposed four distributed processing techniques for classifying URL streams in real-time. The proposed techniques are Independent Stacking, Sequential Stacking, Semi-Sequential Stacking, and Stepwise-Independent Stacking. Our study experimental results showed that Stepwise-Independent Stacking, whose characteristics are similar to those of independent execution and sequential processing, is the best technique for classifying URL streams with the shortest latency.
Å°¿öµå(Keyword) µ¥ÀÌÅÍ ½ºÆ®¸²   µö·¯´× Ã߷Р  ½ºÅÂÅ·   ºÐ»ê 󸮠  ¾ÆÆÄÄ¡ ½ºÅè   data stream   deep learning inference   stacking   distributed processing   Apache Storm                       
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå