[B! æ©Ÿæ¢°å¦ç¿’] HiiragiJPã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

HiiragiJP id:HiiragiJP

æ©Ÿæ¢°å¦ç¿’ã«é–¢ã™ã‚‹HiiragiJPã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (6)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

ã€Žãƒ™ã‚¤ã‚ºæ·±å±¤å¦ç¿’ã€ãŒæœ€é«˜ã™ãŽãŸ - æ—¥å¸¸ã¨é€²æ—
ä»Šå›žã¯æ›¸è©•ã‚¨ãƒ³ãƒˆãƒªãƒ¼ã€‚ ã¡ã‚‡ã†ã©ä»Šæ—¥ã®åˆå‰ä¸ã«é ˆå±±ã•ã‚“ã®ã€Žãƒ™ã‚¤ã‚ºæ·±å±¤å¦ç¿’ã€ã‚’èªã¿çµ‚ãˆãŸã€‚ èªäº†ã€‚ æŽ§ãˆã‚ã«è¨€ã£ã¦ã€ã‚¹ã‚´ã‹ã£ãŸã€‚ ã¾ã˜ã§ãƒœãƒªãƒ¥ãƒ¼ãƒ ãŸã£ã·ã‚Šã§ã‚‚ã®ã™ã”ãèªã¿å¿œãˆã®ã‚ã£ãŸä¸€å†Šã ã£ãŸã¨æ€ã†ã€‚ ãƒ™ã‚¤ã‚ºæ©Ÿæ¢°å¦ç¿’ã«è©³ã—ããªã„äººã§ã‚‚èªã‚ã‚‹ã—(ç°¡å˜ã¨ã¯è¨€ã£ã¦ãªã„)ãƒ›ãƒ³ãƒˆã«å…¨äººé¡žã«ãŠã™ã™ã‚ã—ãŸã„ã€‚ pic.twitter.com/Lbfs6Rr9JMâ€” ã‚³ãƒŸã•ã‚“ (@komi_edtr_1230) January 15, 2020 ã‚‚ã®ã™ã”ãè‰¯ã‹ã£ãŸã®ã§ã“ã“ã§å…¨åŠ›ã§å®£ä¼ã—ã‚ˆã†ã¨æ€ã†ã€‚ æ¦‚è¦ æœ¬æ›¸ã¯ãƒ™ã‚¤ã‚ºçµ±è¨ˆã¨æ·±å±¤å¦ç¿’ã®çµ„ã¿åˆã‚ã›ã«ã¤ã„ã¦è©³èª¬ã—ãŸä¸€å†Šã§ã€é »åº¦è«–ã«åŸºã¥ãç·šå½¢å›žå¸°ã¨ç¢ºçŽ‡åˆ†å¸ƒã®åŸºç¤Žã®è§£èª¬ã‹ã‚‰å§‹ã¾ã‚Šã€ãã“ã‹ã‚‰ç·šå½¢å›žå¸°ã‚„ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ãŒãƒ™ã‚¤ã‚ºçš„ã«ã©ã®ã‚ˆã†ã«èª¬æ˜Žã§ãã‚‹ã‹ã«ã¤ã„ã¦å±•é–‹ã€ãã“ã‹ã‚‰æ·±å±¤å¦ç¿’ã®ãƒ™ã‚¤ã‚ºçš„ãªèª¬æ˜Žã‚’ã—ã¦ã‚¬ã‚¦ã‚¹éŽç¨‹ã¸ã¨ãŸã©ã‚Šç€ãæ§‹æˆã¨ãªã£ã¦ã„ã‚‹ã€‚ æœ¬æ›¸ã®é…åŠ›ã¯ãªã‚“ã¨ã„
HiiragiJP 2020/01/16
æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
æ©Ÿæ¢°å¦ç¿’ãƒ¢ãƒ‡ãƒ«ã‚’ä½œæˆã™ã‚‹ - Training
Microsoft Learn ã§ã¯ã€å¯¾è©±çš„ãªæ–¹æ³•ã§ã€å¾“æ¥ã®æ©Ÿæ¢°å¦ç¿’ã®æ¦‚è¦ã‚’ç†è§£ã™ã‚‹ã“ã¨ãŒã§ãã¾ã™ã€‚ ã“ã‚Œã‚‰ã®ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚° ãƒ‘ã‚¹ã¯ã€ãƒ‡ã‚£ãƒ¼ãƒ— ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°ã®ãƒˆãƒ”ãƒƒã‚¯ã«ç§»è¡Œã™ã‚‹ãŸã‚ã®å„ªã‚ŒãŸåŸºç›¤ã«ã‚‚ãªã‚Šã€å„è‡ªã®ç”Ÿç”£æ€§ã‚’å‘ä¸Šã•ã›ã¾ã™ã€‚ æœ€ã‚‚åŸºæœ¬çš„ãªå¾“æ¥ã®æ©Ÿæ¢°å¦ç¿’ãƒ¢ãƒ‡ãƒ«ã‹ã‚‰ã€æŽ¢ç´¢çš„ãƒ‡ãƒ¼ã‚¿åˆ†æžã‚„ã‚«ã‚¹ã‚¿ãƒžã‚¤ã‚¸ãƒ³ã‚°ã®ã‚¢ãƒ¼ã‚ãƒ†ã‚¯ãƒãƒ£ã¾ã§ã€ãƒ–ãƒ©ã‚¦ã‚¶ãƒ¼ã‚’é›¢ã‚Œã‚‹ã“ã¨ãªãã€æ¦‚å¿µçš„å†…å®¹ã‚„å¯¾è©±åž‹ã® Jupyter Notebook ã‚’ç°¡å˜ã«æŠŠæ¡ã™ã‚‹ã“ã¨ãŒã§ãã¾ã™ã€‚ çŸ¥è˜ã¨èˆˆå‘³ã«å¿œã˜ã¦è‡ªåˆ†ã®ãƒ‘ã‚¹ã‚’é¸æŠžã—ã¦ãã ã•ã„ã€‚ ã‚ªãƒ—ã‚·ãƒ§ãƒ³ 1: å®Œå…¨ãªã‚³ãƒ¼ã‚¹: æ©Ÿæ¢°å¦ç¿’ã®ãŸã‚ã®ãƒ‡ãƒ¼ã‚¿ ã‚µã‚¤ã‚¨ãƒ³ã‚¹ã®åŸºç¤Ž ã»ã¨ã‚“ã©ã®ãƒ¦ãƒ¼ã‚¶ãƒ¼ã«ã¯ã€ã“ã®ãƒ‘ã‚¹ãŒãŠå‹§ã‚ã§ã™ã€‚ ã“ã‚Œã«ã¯ã€æ¦‚å¿µã®ç†è§£ã‚’æœ€å¤§é™ã«é«˜ã‚ã‚‹ã‚«ã‚¹ã‚¿ãƒ ãƒ•ãƒãƒ¼ã‚’å‚™ãˆãŸã€ä»–ã® 2 ã¤ã®ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚° ãƒ‘ã‚¹ã¨åŒã˜ãƒ¢ã‚¸ãƒ¥ãƒ¼ãƒ«ãŒã™ã¹ã¦å«ã¾ã‚Œã¦ã„ã¾ã™ã€‚ åŸºã«ãªã‚‹æ¦‚å¿µã¨ã€æœ€ã‚‚ä¸€èˆ¬çš„ãªæ©Ÿæ¢°å¦ç¿’ãƒ„ãƒ¼ãƒ«ã§ãƒ¢ãƒ‡ãƒ«ã‚’æ§‹
HiiragiJP 2020/01/16
ã‚ã¨ã§èªã‚€

æ©Ÿæ¢°å¦ç¿’

çµ±è¨ˆ
ãƒªãƒ³ã‚¯
2019å¹´ã€ä¿ºã®èªã‚“ã è«–æ–‡50æœ¬å…¨éƒ¨è§£èª¬(ä¿ºçš„ãƒ™ã‚¹ãƒˆ3ä»˜ã) - Qiita
ç¬¬ä¸€ä½ Focal Loss for Dense Object Detection ç‰©ä½“æ¤œçŸ¥ã®ãŸã‚ã®Focal Loss ã“ã‚Œã¯ä¸å‡è¡¡ãƒ‡ãƒ¼ã‚¿ã«å¯¾å‡¦ã™ã‚‹ãŸã‚ã®æå¤±é–¢æ•°Focal Lossã‚’ææ¡ˆã—ãŸè«–æ–‡ãªã®ã§ã™ãŒ, ã¨ã«ã‹ããã®ã‚·ãƒ³ãƒ—ãƒ«ã•ã«ã‚„ã‚‰ã‚Œã¾ã—ãŸ. ç”»åƒã®FLã®å¼ã‚’è¦‹ã¦ãã ã•ã„. ãŸã£ãŸã“ã‚Œã ã‘ã§ã™. ptã¯æ£è§£ãƒ©ãƒ™ãƒ«ã«å¯¾ã™ã‚‹äºˆæ¸¬å€¤ãªã®ã§ã™ãŒ, ptãŒå¤§ãã‘ã‚Œã°å¤§ãã„ã»ã©æå¤±å€¤ã‚’ã—ã£ã‹ã‚ŠæŠ‘ãˆã‚‰ã‚Œã‚‹ã‚ˆã†ã«è¨è¨ˆã•ã‚ŒãŸé–¢æ•°ã¨ãªã£ã¦ã„ã¾ã™. æ£è§£ãƒ©ãƒ™ãƒ«ã«å¯¾ã—ã¦0.6ã¨äºˆæ¸¬ã™ã‚‹ã‚µãƒ³ãƒ—ãƒ«ã‚’å¦ç¿’ã«é‡è¦è¦–ã›ãš, 0.3ã¨ã‹0.1ã¨ã‹äºˆæ¸¬ã—ã¡ã‚ƒã†ã‚µãƒ³ãƒ—ãƒ«ã«ã—ã£ã‹ã‚Šã¨é‡è¦è¦–ã™ã‚‹ã®ã§ã™. è‡ªåˆ†ã‚‚7æœˆãã‚‰ã„ã¾ã§ã¯ä¸å‡è¡¡ãƒ‡ãƒ¼ã‚¿ã«é–¢ã™ã‚‹ç ”ç©¶ã—ã¦ã„ãŸã®ã§ã™ãŒ, è‡ªåˆ†ãŒè€ƒãˆãŸå¤šãã®ã‚¢ã‚¤ãƒ‡ã‚£ã‚¢ãŒçµå±€Focal Lossã®ä¸‹ä½äº’æ›ã«å¸°çµã—ã¦ã—ã¾ã†ã®ã§ã™. ã—ã‹ã‚‚ã“ã®æå¤±é–¢æ•°ã®æ±Žç”¨æ€§ã¯éžå¸¸ã«é«˜ã, ã‚ã‚‰ã‚†ã‚‹ã‚¿ã‚¹ã‚¯ã«å…¥ã‚Œã‚‹ã“ã¨
HiiragiJP 2019/12/30
learning

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
ã€Œã‚€ã—ã‚ã‚¹ã‚ãƒ«ãŒãªã‘ã‚Œã°çµ¦ä¸ŽãŒä¸‹ãŒã‚‹å¯èƒ½æ€§ã€ã€æ±‚ã‚ã‚‰ã‚Œã‚‹ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆäººæã«ãªã‚‹æ–¹æ³•è«–ã‚’ãƒ—ãƒã«ã‚¤ãƒã‹ã‚‰ã„ã‚ã„ã‚èžã„ã¦ã¿ãŸ - GIGAZINE
å…ˆç«¯ITäººæã®ä¸ã§ã‚‚æ€¥å¢—ã™ã‚‹éœ€è¦ã«è¿½ã„ä»˜ã‹ãšã€æ…¢æ€§çš„ãªã€Œå£²ã‚Šæ‰‹å¸‚å ´ã€ã¨ãªã£ã¦ã„ã‚‹ã€Œãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã€ã¨ã„ã†è·ç¨®ã€‚ä¼æ¥ã‹ã‚‰å¼•ãæ‰‹ã‚ã¾ãŸã¨ã„ã‚ã‚Œã‚‹ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆè·ã®å®Ÿæ…‹ã¯ã©ã†ãªã£ã¦ã„ã‚‹ã®ã‹ï¼Ÿã¨ã„ã†ã“ã¨ã§ã€äººæå¤§æ‰‹ãƒ‘ãƒ¼ã‚½ãƒ«ã‚ãƒ£ãƒªã‚¢ã§ç¾å½¹ã®ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã¨ã—ã¦æ´»èºã™ã‚‹é¹¿å†…å¦ã•ã‚“ã«ã€ã„ã‚ã„ã‚ã¨è©±ã‚’èžãã€ã•ã‚‰ã«ä¸–ç•Œæœ€å¤§ç´šã®ã‚ªãƒ³ãƒ©ã‚¤ãƒ³å¦ç¿’æ•™è‚²ãƒ—ãƒ©ãƒƒãƒˆãƒ•ã‚©ãƒ¼ãƒ ã€ŒUdemy(ãƒ¦ãƒ¼ãƒ‡ãƒŸãƒ¼)ã€ã§å½¹ã«ç«‹ã¤è¬›åº§ã¯ã©ã‚Œã‹ã¨ã„ã†ã®ã‚‚æ•™ãˆã¦ã‚‚ã‚‰ã„ã¾ã—ãŸã€‚ ä¸–ç•Œæœ€å¤§ç´šã®ã‚ªãƒ³ãƒ©ã‚¤ãƒ³å¦ç¿’ãƒ—ãƒ©ãƒƒãƒˆãƒ•ã‚©ãƒ¼ãƒ | Udemy ç›®æ¬¡ â—†1ï¼šã€Œãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã€ã¨ã¯ï¼Ÿ â—†2ï¼šã€Œãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã€ãŒå¼•ãæ‰‹ã‚ã¾ãŸãªãƒ¯ã‚± â—†3ï¼šãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ã‚¹ãŒä½¿ãˆãªã„ã¨çµ¦æ–™ãŒæ¸›ã‚‹æ™‚ä»£ãŒæ¥ã‚‹ï¼ï¼Ÿ â—†4ï¼šã€Œãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã€ã«ãªã‚‹ãŸã‚ã®å¦ç¿’æ³•ã¨ã¯ï¼Ÿ â—†5ï¼šãƒ—ãƒãŒãŠã‚¹ã‚¹ãƒ¡ã™ã‚‹Udemyãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ã‚¹è¬›åº§5é¸ ç ”ç©¶è€…ã‹ã‚‰ãƒ“
HiiragiJP 2018/11/21
æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
æ•™å¸«ãªã—å¦ç¿’ã¯æ©Ÿæ¢°ç¿»è¨³ã«é”æ³•ã‚’ã‹ã‘ã‚‹ã‹ï¼Ÿ - ãƒ‡ã‚£ãƒ¼ãƒ—ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°ãƒ–ãƒã‚°
ã¤ã„å…ˆé€±ï¼Œæ©Ÿæ¢°ç¿»è¨³ã§é©šãã¹ãé€²å±•ãŒã‚ã‚Šã¾ã—ãŸï¼Ž æ•™å¸«ãªã—æ©Ÿæ¢°ç¿»è¨³ãŒãƒ¤ãƒã„é€²åŒ–ã‚’é‚ã’ã¦ã¦ã³ã£ãã‚Šã—ãŸï¼ŽãŸã£ãŸåŠå¹´ã§BLEUã‚¹ã‚³ã‚¢ã‚’15ã‹ã‚‰25ã«æ”¹å–„ã—ãŸã®ãƒ–ãƒ¬ã‚¤ã‚¯ã‚¹ãƒ«ãƒ¼ã§ã¯ï¼Ÿhttps://t.co/SVQlYYu2Pt æ•™å¸«ãªã—å¦ç¿’ã§ã“ã®ã‚¯ã‚ªãƒªãƒ†ã‚£ã®æ©Ÿæ¢°ç¿»è¨³ã§ãã‚‹ã®ã¾ã˜ã§æ„Ÿå‹•ã™ã‚‹ã—ï¼Œã¡ã‚‡ã£ã¨èªžã£ã¦ã„ã„ï¼Ÿ pic.twitter.com/fBllGtTkgbâ€” Ryobot | ã‚Šã‚‡ã¼ã£ã¨ (@_Ryobot) 2018å¹´4æœˆ23æ—¥ è¦ç´„ã™ã‚‹ã¨æ•™å¸«ãªã—å¦ç¿’ã§ã‚‚ã²ã¨æ˜”å‰ã®æ•™å¸«ã‚ã‚Šå¦ç¿’ã®æ©Ÿæ¢°ç¿»è¨³ã«åŒ¹æ•µã™ã‚‹æ€§èƒ½ã‚’ç²å¾—ã§ããŸã¨ã„ã†ã®ã§ã™ï¼Žã“ã®è¨˜äº‹ã§ã¯æ©Ÿæ¢°ç¿»è¨³ã‚’çŸ¥ã‚‰ãªã„åˆå¿ƒè€…ã«ã‚‚ã‚ã‹ã‚‹ã‚ˆã†ã«é”æ³•ã®ã‚ˆã†ãªæ•™å¸«ãªã—æ©Ÿæ¢°ç¿»è¨³ã®ä»•çµ„ã¿ã‚’èª¬æ˜Žã—ãŸã„ã¨æ€ã„ã¾ã™ï¼Ž æ•™å¸«ã‚ã‚Šå¦ç¿’ã®é™ç•Œ æ©Ÿæ¢°ç¿»è¨³ã¯ãƒ‡ã‚£ãƒ¼ãƒ—ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°ã‚’é©ç”¨ã™ã‚‹ã“ã¨ã§æ€¥æ¿€ã«é€²æ©ã—ãŸåˆ†é‡Žã®ï¼‘ã¤ã ã¨æ€ã„ã¾ã™ï¼ŽGoogle ç¿»è¨³ã¯ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«æ©Ÿæ¢°ç¿»è¨³ã‚’å°Žå…¥ã™ã‚‹ã“
HiiragiJP 2018/04/28
æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆã®é€†è¥²ã‹ã‚‰ï¼•å¹´å¾Œ | Preferred Research
ç§ãŒ2012å¹´ã«ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆã®é€†è¥²ï¼ˆå½“æ™‚ã®ã‚³ãƒ¡ãƒ³ãƒˆï¼‰ã¨ã„ã†ã®ã‚’ãƒ–ãƒã‚°ã«æ›¸ã„ã¦ã‹ã‚‰ã¡ã‚‡ã†ã©5å¹´ãŒçµŒã¡ã¾ã—ãŸã€‚å½“æ™‚ã¯ã¾ã Deep Learningã¨ã„ã†è¨€è‘‰ãŒåºƒã¾ã£ã¦ãŠã‚‰ãšã€AIã¨ã„ã†è¨€è‘‰ã‚’ä½¿ã†ã®ãŒæ†šã‚‰ã‚Œã‚‹ã‚ˆã†ãªæ™‚ä»£ã§ã—ãŸã€‚ç§é”ãŒã€Preferred Networksï¼ˆPFNï¼‰ã‚’ç«‹ã¡ä¸Šã’IoTã€AIã«ãƒ•ã‚©ãƒ¼ã‚«ã‚¹ã™ã‚‹ã®ã¯ãã‚Œã‹ã‚‰1å¹´åŠå¾Œã®ã“ã¨ã§ã™ã€‚ ã“ã®5å¹´ã‚’æŒ¯ã‚Šè¿”ã‚‹è‰¯ã„ã‚¿ã‚¤ãƒŸãƒ³ã‚°ã ã¨æ€ã†ã®ã§è€ƒãˆã¦ã¿ãŸã„ã¨æ€ã„ã¾ã™ã€‚ 1. Deep Learning Tsunami å¤šãã®åˆ†é‡ŽãŒã“ã®5å¹´é–“ã§Deep Learningã®å¤§ããªå½±éŸ¿ã‚’å—ã‘ã€åˆ†é‡Žç‰¹åŒ–ã®æ‰‹æ³•ãŒDeep Learningãƒ™ãƒ¼ã‚¹ã®æ‰‹æ³•ã«ç½®ãæ›ã‚ã‚‹ã“ã¨ã«ãªã‚Šã¾ã—ãŸã€‚NLPï¼ˆè‡ªç„¶è¨€èªžå‡¦ç†ï¼‰ã®é‡éŽ®ã§ã‚ã‚‹Chris Manningæ•™æŽˆã‚‚NLPã§èµ·ããŸç¾è±¡ã‚’ã€ŒDeep Learning Tsunamiã€[link] ã¨ã‚ˆã³ãã®è¡æ’ƒã®å¤§ãã•ã‚’è¡¨ã—ã¦
HiiragiJP 2017/11/29
æ©Ÿæ¢°å¦ç¿’

AI
ãƒªãƒ³ã‚¯
1