2013-01-01ãã1ã¶æéã®è¨äºä¸è¦§
èªä½ã®ï¼éï¼ç·å½¢åé¡å¨ã®ãªã³ã©ã¤ã³å¦ç¿å¨ã«ï¼ Distributed Training Strategies for the Structured Perceptron (NAACL 2010) ã§ææ¡ããã¦ããåæ£ãªã³ã©ã¤ã³å¦ç¿æ³ (Iterative Parameter Mixing) ãå®è£ ãã¦ã¿ãï¼ æ©æ¢°å¦ç¿ à MapReduce - ny23ã®æ¥è¨ ã§â¦
大æçç²åº 2012 - ny23ã®æ¥è¨ ããã®ç¶ãï¼ [01/23] å¹´æ«ã«200gè±ãè²·ã£ã¦ï¼1/6 ã¯è±ãæ®ã£ã¦ããã®ã§çç²ã ãï¼ãã®æ¬¡ã®é±ï¼å ã é±ï¼ã¯èªåãç¨äºããããä¼ã¿ï¼å é±ã¯å¤§æããããä¼ã¿ã§ä»å¹´åãã¦ã®è±è³¼å ¥ï¼çç²è±ããªãã®ã§æ¬å½ã«å°ã£ã¦ããï¼æåã¯æ¬â¦
æå°ãã¦ããï¼ãï¼å¦çãã¡ï¼åæ¥çå«ãï¼ãï¼ç«ã¦ç¶ãã«è«æèªãç 究ä¼ã§è«æè³ãããã£ããããï¼è¯ããã¨ãéãªããã®ã ï¼ã©ã¡ããå ±èè ã¨ãã¦ãã£ããé¢ãã£ãç 究ãªã®ã ãã©ï¼ç¹ã«è«æèªã®æ¹ã¯ç¬¬äºèè ã ã£ãã®ã§æå¾ã¾ã§ã³ã¡ã³ããå ¥ãããã¨ãããâ¦
ç·å½¢åé¡å¨ã®ã©ã¤ãã©ãªã«å梱ãã¦å ¬éããåçããã«é åã«åé¤æ©è½ã¨ç»é²ãã¼ï¼å¤ã®ãã³ãæ©è½ã追å å®è£ ãã¦ï¼å¥ããã±ã¼ã¸ã¨ãã¦å ¬éããï¼æ¨æºçãªéçããã«é åã®ã©ã¤ãã©ãªã®ã¤ã³ã¿ã¼ãã§ã¹ã«å ãã¦ï¼è¿½å ï¼åé¤ï¼ç»é²ãã¼ã®ãã³ããå®è£ ãã¦ããï¼â¦