ãã©ã³ã¯ãã³ã®ç»ååé¡ã³ã³ããéå¬ããã¦ãã¾ã!
Kaggleã§ãã©ã³ã¯ãã³ã®ç»ååé¡ã¢ãã«ã®è¯ãã競ãã³ã³ããéå¬ããã¦ãã¾ãã
Description - National Data Science Bowl | Kaggle
Computer VisionãDeep Learningã«èå³ã®ããæ¹ãèªåã®åã試ãã¦ã¿ããæ¹ãæé ãªãµã³ãããã¯ã欲ããæ¹ãªã©ã¯åå ããã¦ã¿ã¦ã¯ãããã§ããããã
- æ¦è¦(è¶ è¨³)
- æµ·ã®ç¶æ ããã©ã³ã¯ãã³ã®ç¨®é¡ãæ°ã®åå¸ãªã©ãã調ã¹ããããã©ã³ã¯ãã³ã®èå¥ã人éãè¡ãã®ã¯å¤§å¤ãªã®ã§æ°´ä¸ã«ã¡ã©ã§æ®å½±ãããã©ã³ã¯ãã³ã®ç»åãã種é¡ãäºæ¸¬ããã¢ãã«ãä½ã£ã¦ãããããã
- æé
- 2014/12/15ã2015/3/16
- è³éç·é¡
- $175,000 (14/12/20: ç´2090ä¸å)
- ãã¼ã¿
- å¦ç¿ç»å: ç´3ä¸æ (ã°ã¬ã¼ã¹ã±ã¼ã«/æ§ã ãªãµã¤ãº/ä¸åä¸), 対象ã¯ã©ã¹: 121種é¡(å¤ã¯ã©ã¹åé¡/確çåºå/multi-class logarithmic lossã§è©ä¾¡), ãã¹ãç»å: ç´13ä¸æ (ãã¼ã対çã§è©ä¾¡ã«ä½¿ç¨ãããªããã¼ã¿ãããããå«ã)
詳細ã¯ããµã¤ããåç
§ãã ããã
注æãããã¨ã¨ãã¦ããã¹ããã¼ã¿ã使ã£ã¦ã®åæ師ããå¦ç¿ãæ師ãªãç¹å¾´å¦ç¿ã¯è¨±å¯ããã¦ãã¾ãããå¤é¨ãã¼ã¿ã使ããã¨ã¯ç¦æ¢ããã¦ãã¾ãã(Caffeã®pre-trained ImageNet modelãªã©ã¯ä½¿ç¨ä¸å¯)
è³éãå¤ããã¨ããããããã人ãæ¥ãã¨æããããã¾ã ãã¾ãããªããããªã®ã§ã宣ä¼ãæ¸ãã¾ãããå½ç¶åãåå ãã¦ãã¾ãã