[B! igo] nfunatoã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

nfunato id:nfunato

igoã«é–¢ã™ã‚‹nfunatoã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (24)

${{author_name}}$

{{{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}}

å¤¢ã ã‘ã§æŒ‘ã‚“ã ã€Œå€‹äººã‚²ãƒ¼ãƒ é–‹ç™ºã€ç‹¬ç«‹ã—ãŸã‚‰å¹´åŽ15ä¸‡å††ã§ãƒ‡ãƒƒãƒ‰ã‚¨ãƒ³ãƒ‰é–“è¿‘ãªç”·ã¨ã€ä»®æƒ³é€šè²¨ã‚¢ãƒ—ãƒªã®åŽç›ŠãŒã€Œæœˆ30ä¸‡å††ã‹ã‚‰æœˆæ•°åƒå††ã€ã«æ€¥è½ã—ãŸè©±
ã‚¢ãƒ—ãƒªãƒžãƒ¼ã‚±ãƒ†ã‚£ãƒ³ã‚°ç ”ç©¶æ‰€ > ã‚¢ãƒ—ãƒªé–‹ç™º > å¤¢ã ã‘ã§æŒ‘ã‚“ã ã€Œå€‹äººã‚²ãƒ¼ãƒ é–‹ç™ºã€ç‹¬ç«‹ã—ãŸã‚‰å¹´åŽ15ä¸‡å††ã§ãƒ‡ãƒƒãƒ‰ã‚¨ãƒ³ãƒ‰é–“è¿‘ãªç”·ã¨ã€ä»®æƒ³é€šè²¨ã‚¢ãƒ—ãƒªã®åŽç›ŠãŒã€Œæœˆ30ä¸‡å††ã‹ã‚‰æœˆæ•°åƒå††ã€ã«æ€¥è½ã—ãŸè©± 6åã®ã‚¢ãƒ—ãƒªé–‹ç™ºè€…ã‚’å–æã—ã¾ã—ãŸã€‚ã€Œå€‹äººé–‹ç™ºè€…ç‰¹é›†2018ã€ã®ç¬¬å…å›žï¼ˆæœ€çµ‚å›žï¼‰ã§ã™ã€‚ <ç›®æ¬¡> ï¼‘ã€ãƒ‡ã‚¹ãƒžãƒ¼ãƒã‚’è¶…ãˆã¦å¤¢ã ã‘ã§ç‹¬ç«‹ã—ãŸã‚‰å¹´åŽ15ä¸‡å†† ï¼’ã€ä»®æƒ³é€šè²¨ã‚¢ãƒ—ãƒªã®ã‚¢ãƒ•ã‚£ãƒªã‚¨ã‚¤ãƒˆæœˆ30ä¸‡å††ãŒæœˆæ•°åƒå††ã« ï¼“ã€ãƒ‰ãƒãƒ¼ãƒ³é£›è¡Œãƒžãƒƒãƒ—é–‹ç™ºè€…ã®ãƒ‹ãƒƒãƒã‚¢ãƒ—ãƒªã¤ãã‚‹æ¥½ã—ã• ï¼”ã€300ä¸‡DLã®ãƒ›ãƒ©ãƒ¼ã‚²ãƒ¼é–‹ç™ºè€…ãŒèªžã‚‹ãƒ›ãƒ©ãƒ¼ã®åºƒãŒã‚Šæ–¹ ï¼•ã€å‰¯æ¥é–‹ç™ºã®çž‘æƒ³ã‚¿ã‚¤ãƒžãƒ¼ã‚¢ãƒ—ãƒªãŒæœˆ2ä¸‡å††ã®åºƒå‘ŠåŽå…¥ ï¼–ã€å›²ç¢ã‚¢ãƒ—ãƒªãŒåŽç›Š80ä¸‡å††ã€‚65æ³ä»¥ä¸Šãƒ¦ãƒ¼ã‚¶ãƒ¼ãŒå¤šã„ ï¼‘ã€ã‚¢ãƒ—ãƒªã§ç‹¬ç«‹ã€Œå¤¢ã ã‘ã€ã§æŒ‘æˆ¦ã—ãŸã‚‰å¹´åŽ15ä¸‡å†† â€»å€‹äººé–‹ç™ºè€…ã®ãµã‚Šãµã‚‰ã•ã‚“ï¼ˆ30ä»£ åŸ¼çŽ‰çœŒåœ¨ä½ï¼‰ ã‚‚ã¨ã‚‚ã¨ã€Œé–‹ç™ºè€…ã«ãªã‚‹ã¾ã§ã€ã¯ä½•ã‚’ã•ã‚Œã¦ã„ãŸã‚“ã§ã™ã‹ã€‚ ã‚‚ã¨ã‚‚ã¨ã¯ã€ã‚½ãƒ•ãƒˆã‚¦
nfunato 2018/12/18
business

igo
ãƒªãƒ³ã‚¯
ãƒ¢ãƒ³ãƒ†ã‚«ãƒ«ãƒæœ¨æŽ¢ç´¢æ³• (Commutative Weblog)
ãƒ¢ãƒ³ãƒ†ã‚«ãƒ«ãƒæœ¨æŽ¢ç´¢ã¨ã¯ã€èžããªã‚Œãªã„ç”¨èªžã§ã‚ã‚‹ãŒã€ãªã‚“ã§ã‚‚æœ€è¿‘ã€ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿å›²ç¢ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã®åˆ†é‡Žã§ä½¿ã‚ã‚Œã¦ã„ã‚‹æ‰‹æ³•ã‚‰ã—ã„ã€‚æƒ…å ±å‡¦ç†ã¨ã„ã†é›‘èªŒã®ï¼’ï¼ï¼ï¼˜å¹´ï¼–æœˆå·ã®ã€ç¾Žæ·»ä¸€æ¨¹æ°ã®è¨˜äº‹ã«å¾“ã„ã€ç´¹ä»‹ã—ã¦ã¿ã‚‹ã€‚ æœ€è¿‘ã€ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿å°†æ£‹ã®é€²å¢ƒãŒè‘—ã—ã„ã€‚æ¸¡è¾ºç«œçŽ‹å¯¾ãƒœãƒŠãƒ³ã‚¶ã®å¯¾æˆ¦ãŒç¤ºã™ã‚ˆã†ã«ã€ãƒˆãƒƒãƒ—ãƒ—ãƒã§ã‚‚ã€æ°—åˆã‚’å…¥ã‚Œã¦ã‹ã‹ã‚‰ãªã„ã¨ç°¡å˜ã«ã¯å‹ã¦ãªã„ã»ã©ã§ã‚ã‚‹ã€‚ ä¸€æ–¹ã€ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿å›²ç¢ã§ã¯ã©ã†ã‹ã€‚ã¡ã‚‡ã£ã¨å‰ã¾ã§ã€ã‚„ã£ã¨ã‚¢ãƒžãƒãƒ¥ã‚¢ä¸‰ç´šãã‚‰ã„ã‹ã‚‚ã€ã¨è¨€ã‚ã‚ŒãŸã€‚ã©ã†ã—ã¦ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿å°†æ£‹ã«æ¯”ã¹ã¦å¼±ã„ã®ã‹ã€‚ä¸€ç•ªåˆ†ã‹ã‚Šã‚„ã™ã„èª¬æ˜Žã¯ã€ï¼‘ï¼™è·¯ç›¤ã ã¨ã€æŽ¢ç´¢ç©ºé–“ãŒè†¨å¤§ã™ãŽã‚‹ã€ã¨ã„ã†ã‚‚ã®ã§ã‚ã‚‹ã€‚ã—ã‹ã—ã€æŽ¢ç´¢ç©ºé–“ã ã‘ã®å•é¡Œã§ã‚‚ãªã„ã‚ˆã†ã§ã‚ã‚‹ã€‚ä¾‹ãˆã°ã€ï¼™è·¯ç›¤ã ã¨ã€æŽ¢ç´¢ç©ºé–“ã¯ã€ãƒã‚§ã‚¹ã‚ˆã‚Šã‚‚ç‹ã„ã€‚ã—ã‹ã—ã€æ—§æ¥ã®ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ ã§ã¯ã€ï¼™è·¯ç›¤ã§ã‚‚ã€ã‚¢ãƒžãƒãƒ¥ã‚¢åˆæ®µã«å‹ã¦ãªã„ã€‚ã©ã†ã‚‚ã€æŽ¢ç´¢ç©ºé–“ã ã‘ã§ã¯ãªãã€è©•ä¾¡é–¢æ•°ã«å•é¡ŒãŒã‚ã‚‹ã‚ˆã†ãªã®ã§ã‚ã‚‹ã€‚
nfunato 2018/12/08
game

igo
ãƒªãƒ³ã‚¯
ELF/README.rst at master Â· pytorch/ELF
nfunato 2018/05/03
igo

facebook
ãƒªãƒ³ã‚¯
Performance Trends in AI
Epistemic Status: moderately confident Edit To Add:Â Itâ€™s been brought to my attention that I was wrong to claim that progress in image recognition is â€œslowing downâ€. As classification accuracy approaches 100%, obviously improvements in raw scores will be smaller, by necessity, since accuracy canâ€™tÂ exceed 100%. If you look at negative log error rates rather than raw accuracy scores, improvement in
nfunato 2017/10/28
"Deep learning hasnâ€™t really proven itself better than trendline in strategy games (Go and chess) or in natural language tasks. ... AlphaGo seems relevantly different mainly because most of its capacity was precomputed. â€

deeplearning

nlp

igo
ãƒªãƒ³ã‚¯
äººé–“VSã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ã‚ªã‚»ãƒã€€è¡æ’ƒã®6æˆ¦å…¨æ•—ã‹ã‚‰20å¹´ã€å…ƒä¸–ç•Œãƒãƒ£ãƒ³ãƒ”ã‚ªãƒ³æ‘ä¸Šå¥ã•ã‚“ã«èžã„ãŸã€Œè² ã‘ãŸå¾Œã«è¦‹ãˆã¦ããŸã‚‚ã®ã€ (1/3) - ãã¨ã‚‰ã¼
2017å¹´ã€äººé–“ã¯2ã¤ã®çŸ¥çš„ã‚²ãƒ¼ãƒ ã§ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ã«æ±ºå®šçš„ãªæ•—åŒ—ã‚’å–«ã—ã¾ã—ãŸã€‚ å›²ç¢ã®ä¸–ç•Œãƒ¬ãƒ¼ãƒ†ã‚£ãƒ³ã‚°1ä½ã®æŸ¯æ½”ï¼ˆã‚«ãƒ»ã‚±ãƒ„ï¼‰ä¹æ®µãŒã€ç±³Googleå‚˜ä¸‹DeepMindã®å›²ç¢AIã€ŒAlpha Goã€ã¨ã®3ç•ªå‹è² ã§3æˆ¦å…¨æ•—ï¼ˆé–¢é€£è¨˜äº‹ï¼‰ã€‚ ãƒœãƒ¼ãƒ‰ã‚²ãƒ¼ãƒ æœ€å¾Œã®ç ¦ã¨ã„ã‚ã‚ŒãŸå›²ç¢ã•ãˆCOMã«æ•—åŒ— ã•ã‚‰ã«å°†æ£‹ã®ä½è—¤å¤©å½¦åäººãŒã€ç¬¬2æœŸé›»çŽ‹æˆ¦äºŒç•ªå‹è² ã§å°†æ£‹AIã€ŒPONANZAï¼ˆãƒãƒŠãƒ³ã‚¶ï¼‰ã€ã«ã€å…ˆæ‰‹ç•ªãƒ»å¾Œæ‰‹ç•ªã¨ã‚‚ã«æ•—ã‚Œã¾ã—ãŸï¼ˆé–¢é€£è¨˜äº‹ï¼‰ã€‚ æŠ•äº†ç›´å‰ã€å¤©ã‚’ä»°ãåäººï¼ˆãƒ‹ã‚³ãƒ‹ã‚³ç”Ÿæ”¾é€ã‚ˆã‚Šï¼‰ æ€¥æ¿€ã«é€²æ©ã™ã‚‹AIã«ã‚ˆã‚Šç”Ÿæ´»ãŒæ¿€å¤‰ã™ã‚‹ã¨ã„ã‚ã‚Œã‚‹21ä¸–ç´€ã€‚2045å¹´ã€ã‚ã‚‹ã„ã¯ãã‚Œã‚’ä¸Šå›žã‚‹é€Ÿåº¦ã§ã€äººé–“ã®çŸ¥èƒ½ã‚’AIãŒæ±ºå®šçš„ã«ä¸Šå›žã‚‹ã€Œã‚·ãƒ³ã‚®ãƒ¥ãƒ©ãƒªãƒ†ã‚£ã€ãŒæ¥ã‚‹ã¨ã‚‚ã„ã‚ã‚Œã€äººé–“ã®å˜åœ¨ä¾¡å€¤ã™ã‚‰å•ã‚ã‚Œå§‹ã‚ã¦ã„ã‚‹æ˜¨ä»Šã«ãŠã„ã¦ã€è¡æ’ƒçš„ãªå‡ºæ¥äº‹ã§ã—ãŸã€‚ ãã®ã¡ã‚‡ã†ã©20å¹´å‰ã®1997å¹´ã€‚å›²ç¢ã¨å°†æ£‹ã®ã‚ˆã†ã«ã€äººé–“ã«ã¨ã£ã¦éžå¸¸ã«ãƒãƒ”ãƒ¥ãƒ©ãƒ¼ãªçŸ¥çš„ã‚²ãƒ¼ãƒ
nfunato 2017/10/21
cshogi

igo
ãƒªãƒ³ã‚¯
Zeroå¯¾Masteræ£‹èœæ„Ÿæƒ³01
ä»Šå›žå…¬é–‹ã•ã‚ŒãŸAlpha GoZeroã®æ£‹èœã®ã†ã¡ã€å¯¾Masteræˆ¦ã‚’å–ã‚Šä¸Šã’ã¦ã¿ãŸã„ã€‚ ç§ã¯ã¤ã„å…ˆæ—¥ã€Masterã®60å±€ã«ã¤ã„ã¦ã®æ„Ÿæƒ³ã‚’æ›¸ãä¸Šã’ãŸã°ã‹ã‚Šã ã€‚ ãã®é–“ã€è¡æ’ƒçš„ãªAlpha Goè‡ªå·±å¯¾æˆ¦50å±€ã®æ£‹èœç™ºè¡¨ãŒã‚ã£ãŸãŒã€Masterã®60å±€ãŒçµ‚ã‚ã‚‹ã¾ã§ã¯ã€ã‚ãˆã¦è¦‹ãªã„ã‚ˆã†ã«ã—ã¦ã„ãŸã€‚ AIã®é€²åŒ–ã®é€Ÿã•ã«ã¯é©šãã°ã‹ã‚Šã ã€‚ã“ã®ã‚¹ãƒ”ãƒ¼ãƒ‰ã¯ã€èƒŒæ™¯ã«ã‚ã‚‹å·¨å¤§ãªåŠ›ã«å½±éŸ¿ã•ã‚Œã¦ã„ã‚‹ã¨ã¯è¨€ãˆã€åºç›¤ã®ä¸Šæ‰‹ã•ã¨æˆ¦é—˜åŠ›ã€çµ‚ç›¤åŠ›ã€å…¨ã¦ã‚’æ¥µã‚ãŸã¨æ€ã‚ã‚ŒãŸMasterã‚’åœ§å€’ã™ã‚‹ã‚‚ã®ãŒã€ä»Šå‡ºã¦ãã‚‹ã¨ã¯ã€æœ¬å½“ã«å‡„ã™ãŽã‚‹ã€‚ å¸ƒçŸ³ã«ã¤ã„ã¦ã‚‚ã€ç‹¬è‡ªã®æ–°ã—ã„ã‚¢ã‚¤ãƒ‡ã‚¢ãŒè¦‹ã‚‰ã‚Œã‚‹ãŒã€ãã“ã§MasterãŒã²ã‘ã‚’å–ã£ã¦ã„ã‚‹è¨³ã§ã¯ãªã•ãã†ã ã€‚ ä½•ã¨è¨€ã£ã¦ã‚‚ä¸ç›¤ä»¥é™ã®æã‚ã—ã„åŠ›ã§ã€çªãæ”¾ã—ã¦ã„ã‚‹ï¼ˆã¨æ„Ÿã˜ã‚‰ã‚Œã‚‹ï¼‰ã€‚å·¦è¾ºã¸ã®ä¾µå…¥ã€ä¸å¤®ã§ã®ãƒ¨ã‚»ã€‚æƒ³åƒã‚‚ã¤ã‹ãªã„è†¨å¤§ãªé‡ã®ãƒ¨ãƒŸã‚’ã“ãªã—ã¦ã„ã‚‹äº‹ã ã‚ã†ã€‚ Zeroã¯ã€äººé–“ãŒä½•åƒå¹´ã‚‚ç©ã¿é‡ããŸè‰¯ã„æ‰‹
nfunato 2017/10/20
igo
ãƒªãƒ³ã‚¯
GitHub - ymgaq/AQ: A Go game engine. Download here: http://github.com/ymgaq/AQ/releases
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session. Dismiss alert
nfunato 2017/09/18
igo

ai

github
ãƒªãƒ³ã‚¯
New community features for Google Chat and an update on Currents
nfunato 2017/07/31
igo

ai

deeplearning
ãƒªãƒ³ã‚¯
é‡‘æ²¢å…„å¼Ÿ
nfunato 2017/07/09
igo
ãƒªãƒ³ã‚¯
è¡æ’ƒï¼AlphaGoã®è‡ªå·±å¯¾æˆ¦â”‚å›²ç¢ã‚²ãƒ¼ãƒ ã®ãƒ‘ãƒ³ãƒ€ãƒãƒƒãƒˆ
Alpha GoãŒå†ã³å…¬ã®å ´ã«å§¿ã‚’ç¾ã—ãŸã€Œæœªæ¥ã®ç¢ã‚µãƒŸãƒƒãƒˆã€ã¯ã€æŸ¯æ½”ä¹æ®µã¨ã®ä¸‰ç•ªå‹è² ã€ç›¸è«‡ç¢ã¨ã‚‚ã«Alpha GoãŒå®‰å®šã—ãŸå¼·ã•ã‚’è¦‹ã›ã¦ã®å‹åˆ©ã¨ãªã‚Šã¾ã—ãŸã€‚ç‰¹ã«é¢ç™½ã‹ã£ãŸã®ã¯å¤åŠ›ä¹æ®µï¼†Alpha Goå¯¾é€£ç¬‘ä¸ƒæ®µAlpha Goã®ãƒšã‚¢ç¢å¯¾æ±ºã§ã—ãŸã€‚äººé¡žã¨AIåˆã®ç«¶æ¼”ã¨ãªã£ãŸã‚ã‘ã§ã™ãŒã€ã‚¹ãƒªãƒªãƒ³ã‚°ãªå†…å®¹ã¨ãªã‚Šã€ã“ã‚Œã‹ã‚‰ã®AIã¨ã®é–¢ã‚ã‚Šæ–¹ã‚’è€ƒãˆã‚‹ã†ãˆã§éžå¸¸ã«æœ‰æ„ç¾©ã ã£ãŸã®ã§ã¯ãªã„ã‹ã¨æ€ã„ã¾ã™ã€‚ä¸€æ™‚ã¯å¤åŠ›ä¹æ®µãƒšã‚¢ãŒå‹çŽ‡75ï¼…ã¾ã§å„ªå‹¢ã«ãªã‚ŠãªãŒã‚‰é€†è»¢ã‚’è¨±ã—ãŸã®ã¯ã€ç§ã«ã¨ã£ã¦ã‚‚é©šãã§ã—ãŸã€‚ ä¸€ä½“Alpha Goã¯ã©ã‚Œãã‚‰ã„å¼·ã„ã®ã‹ï¼Ÿç¾åœ¨ã®æ‰€ã€Alpha Goã¯å¼·ããªã‚Šã™ãŽã¦ã€å®ŸåŠ›ã‚’æ£ç¢ºã«æ¸¬ã‚‹æ‰‹æ®µã¯ãªã„ã‚ˆã†ã«è¦‹ãˆã¾ã™ã€‚ã—ã‹ã—ã€Œæœªæ¥ã®ç¢ã‚µãƒŸãƒƒãƒˆã€é–‰ä¼šå¼ã§DeepMindãƒ‡ãƒŸã‚¹ãƒ»ãƒã‚µãƒ“ã‚¹æ°ã‚ˆã‚Šè¡æ’ƒã®ç™ºè¡¨ãŒã‚ã‚Šã¾ã—ãŸã€‚ãªã‚“ã¨Alpha Goã®è‡ªå·±å¯¾æˆ¦50å±€ã‚’å…¬é–‹ã™ã‚‹ã¨ã„ã†ã®ã§ã™ã€‚ã“ã‚Œã‚‰ã¯ãã¡ã‚“ã¨è€ƒæ…®æ™‚é–“ã‚’å–ã£ã¦æ‰“ãŸã‚ŒãŸå¯¾å±€
nfunato 2017/06/12
DeepMind

alphago

igo
ãƒªãƒ³ã‚¯
GoogleãŒå‡ºã—ãŸå›²ç¢ã‚½ãƒ•ãƒˆã€ŒAlphaGoã€ã®è«–æ–‡ã‚’ç¿»è¨³ã—ã¦è§£èª¬ã—ã¦ã¿ã‚‹ã€‚ - 7rpnâ€™s blog: ã†ã‚ã‚ã‚ã‚ã‚ãªæ—¥å¸¸
å°±æ´»ã‚‚ç„¡äº‹çµ‚ã‚ã£ãŸã®ã§ï¼Œä¸€ç•ªã‚„ã‚ŠãŸã‹ã£ãŸAlpha Goã®è«–æ–‡ã‚’ç¿»è¨³ã—ã¾ã—ãŸã€‚ ã”å˜çŸ¥ã®é€šã‚Šï¼Œé•·ã‚‰ãä¸–ç•Œæœ€å¼·ã ã£ãŸå›²ç¢æ£‹å£«ã‚¤ãƒ»ã‚»ãƒ‰ãƒ«ä¹æ®µã‚’ç ´ã£ãŸGoogleã®å›²ç¢ãƒ—ãƒã‚°ãƒ©ãƒ ã§ã™ã€‚ è«–æ–‡ã®å†…å®¹ã«è§¦ã‚Œã¤ã¤ä½•ã¨ãªãè§£èª¬å…¥ã‚Œã¦ã„ããŸã„ã¨æ€ã„ã¾ã™ã€‚ãªã‚‹ã¹ãå›²ç¢ã‚„DeepLearningã‚’çŸ¥ã‚‰ãªã„åˆå¿ƒè€…ã®äººã¨ã‹ã§ã‚‚ä»•çµ„ã¿ã‚’ç†è§£ã§ãã‚‹ã‚ˆã†ã«åˆ†ã‹ã‚Šã‚„ã™ãæ›¸ã„ã¦ã„ã‘ã‚Œã°ã„ã„ãªã¨æ€ã„ã¾ã™ã€‚ åŽŸé¡Œã¯"Mastering the game of Go with deep neural networks and tree search"ã€‚ ã¨ã‚Šã‚ãˆãšæœ€åˆã®è¦ç´„ã®è¨³ã‹ã‚‰ã€‚ è¬Žã®å˜èªžã¨ã‹ã¯å¾Œã‹ã‚‰èª¬æ˜Žå…¥ã‚Œã‚‹ã®ã§ï¼Œã•ã‚‰ã£ã¨æµã—èªã¿ã—ã¦ã„ãŸã ã‘ã‚Œã°ã€‚ å›²ç¢ã¯ã“ã‚Œã¾ã§AIã«ã¨ã£ã¦ã¨ã¦ã‚‚é›£ã—ã„ã‚²ãƒ¼ãƒ ã ã¨ã¿ãªã•ã‚Œã¦ããŸã€‚ãã‚Œã¯æŽ¢ç´¢ç¯„å›²ãŒã¨ã¦ã‚‚åºƒã„ã“ã¨ã¨ï¼Œç›¤é¢ã®è©•ä¾¡ãŒé›£ã—ã„ãŸã‚ã€‚ ã“ã®è«–æ–‡ã§ã¯ï¼Œã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ã‚’ç”¨ã„ãŸå›²ç¢ã®æ–°ã—ã„ã‚¢ãƒ—ãƒãƒ¼
nfunato 2017/01/04
igo

ai
ãƒªãƒ³ã‚¯
ãªãœã€Œå›²ç¢ã€ã ã£ãŸã®ã‹ã€‚ãªãœã€Œ10å¹´ã‹ã‹ã‚‹ã€ã¨è¨€ã‚ã‚Œã¦ã„ãŸã®ã‹â”€â”€AlphaGoå‰æ—¥èš
nfunato 2016/11/20
cshogi

igo
ãƒªãƒ³ã‚¯
å›²ç¢AIã€ŒAlphaGoã€ã¯ä½•ãŒã™ã”ã„ã®ï¼Ÿ æ·±å±¤å¦ç¿’Ã—å¼·åŒ–å¦ç¿’ãŒäººå·¥çŸ¥èƒ½ã«ä¸ŽãˆãŸã‚¤ãƒ³ãƒ‘ã‚¯ãƒˆ
ã€ŽMarkeZineã€ãŒä¸»å‚¬ã™ã‚‹ãƒžãƒ¼ã‚±ãƒ†ã‚£ãƒ³ã‚°ãƒ»ã‚¤ãƒ™ãƒ³ãƒˆã€ŽMarkeZine Dayã€ã€ŽMarkeZine Academyã€ã€ŽMarkeZine ãƒ—ãƒ¬ãƒŸã‚¢ãƒ ã‚»ãƒŸãƒŠãƒ¼ã€ã® æœ€æ–°æƒ…å ±ã‚’ã¯ã˜ã‚ã€æ§˜ã€…ãªã‚¤ãƒ™ãƒ³ãƒˆæƒ…å ±ã‚’ã¾ã¨ã‚ã¦ã”ç´¹ä»‹ã—ã¾ã™ã€‚ MarkeZine Day
nfunato 2016/10/21
deeplearning

igo
ãƒªãƒ³ã‚¯
Deep Learningãƒ©ã‚¤ãƒ–ãƒ©ãƒª{mxnet}ã®Rç‰ˆã§Convolutional Neural Networkã‚’ã‚µã‚¯ãƒƒã¨è©¦ã—ã¦ã¿ãŸï¼ˆè¿½è¨˜3ä»¶ã‚ã‚Šï¼‰ - æ¸‹è°·é§…å‰ã§åƒããƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã®ãƒ–ãƒã‚°
For non-native Japanese speakers: English version is below. ã¡ã‚‡ã£ã¨å‰ã‹ã‚‰è‰²ã€…ãªã¨ã“ã‚ã§ã¡ã‚‰ã»ã‚‰åå‰ã‚’èžããªãã¨æ€ã£ã¦ã„ãŸMXnetã€‚ã©ã†ã‚„ã‚‰Kagglerã®é–“ã§ã¯æ€¥é€Ÿã«äººæ°—ãŒé«˜ã¾ã£ã¦ã„ã‚‹ã‚ˆã†ã§ã€æœ€è¿‘ã«ãªã£ã¦ã ã„ã¶ãƒã‚°ãƒ•ã‚£ãƒƒã‚¯ã‚¹ãŒé€²ã‚“ã ã‚‰ã—ã„ã¨ã„ã†ã®ã§åƒ•ã‚‚ã‚¤ãƒ³ã‚¹ãƒˆãƒ¼ãƒ«ã—ã¦ã¿ã‚‹ã“ã¨ã«ã—ã¾ã—ãŸã€‚ ã‚‚ã†ã“ã‚Œã¯ã”è¦§ã®é€šã‚Šã§ã€ã‚ã®Xgboostã‚‚é…å¸ƒã—ã¦ã„ã‚‹DMLCãŒå‡ºã—ãŸã‚‚ã®ã§ã™ã€‚æ—¢ã«ã‚½ãƒ¼ã‚¹ã‚‚ãƒ‰ã‚ãƒ¥ãƒ¡ãƒ³ãƒˆã‚‚ã‹ãªã‚Šæ•´å‚™ãŒé€²ã‚“ã§ã„ã¦ã€ImageNetã®å¦ç¿’æ¸ˆã¿ãƒ¢ãƒ‡ãƒ«ã‚‚é…å¸ƒã•ã‚Œã¦ã„ã‚‹ã‚ˆã†ã§ã™ã€‚ã¨ã„ã†ã“ã¨ã§ã€æ—©é€Ÿã¡ã‚‡ã£ã¨è§¦ã£ã¦ã¿ã‚ˆã†ã¨æ€ã„ã¾ã™ã€‚ Convolutional Neural Network (CNN)ã¨ã¯ è¨€ã†ã¾ã§ã‚‚ãªã„ã“ã¨ã ã¨ã¯æ€ã„ã¾ã™ãŒã€Convolutional Neural Network (CNN)ã¨ã„ã†æ‰‹æ³•è‡ªä½“ã¯Deep
nfunato 2016/06/06
deeplearning

igo
ãƒªãƒ³ã‚¯
Monte-Carlo Tree Search: A New Framework for Game AI
nfunato 2016/05/18
igo
ãƒªãƒ³ã‚¯
What programming language did Google use to create AlphaGo?
nfunato 2016/05/18
c++

lua

igo
ãƒªãƒ³ã‚¯
cakesï¼ˆã‚±ã‚¤ã‚¯ã‚¹ï¼‰
cakesã¯2022å¹´8æœˆ31æ—¥ã«çµ‚äº†ã„ãŸã—ã¾ã—ãŸã€‚ 10å¹´é–“ã®é•·ãã«ã‚ãŸã‚Šã€ã”æ„›èªã‚ã‚ŠãŒã¨ã†ã”ã–ã„ã¾ã—ãŸã€‚ 2022å¹´9æœˆ1æ—¥
nfunato 2016/04/06
cshogi

igo
ãƒªãƒ³ã‚¯
ã€Œã‚¢ãƒ«ãƒ•ã‚¡ç¢ã€ãŒä¸–ç•Œãƒˆãƒƒãƒ—ç´šæ£‹å£«ã«å‹ã¡è¶Šã—ã€€äººé–“ã®ã€Œå¤§å±€è¦³ã€è¶…ãˆã‚‹äººå·¥çŸ¥èƒ½ï¼ˆæœ¨æ‘æ£äººï¼‰ - ã‚¨ã‚ã‚¹ãƒ‘ãƒ¼ãƒˆ - Yahoo!ãƒ‹ãƒ¥ãƒ¼ã‚¹
ï¼•ç•ªå‹è² ã§ï¼“é€£å‹ãƒãƒ³ãƒ‰ãƒ³ã«ã‚ã‚‹äººå·¥çŸ¥èƒ½é–‹ç™ºä¼šç¤¾ã€Œã‚°ãƒ¼ã‚°ãƒ«ãƒ»ãƒ‡ã‚£ãƒ¼ãƒ—ãƒžã‚¤ãƒ³ãƒ‰ã€ãŒé–‹ç™ºã—ãŸãƒ—ãƒã‚°ãƒ©ãƒ ã€Œã‚¢ãƒ«ãƒ•ã‚¡ç¢ã€ãŒ12æ—¥ã€éŸ“å›½ã®ãƒ—ãƒæ£‹å£«ã§ä¸–ç•Œãƒˆãƒƒãƒ—ãƒ¬ãƒ™ãƒ«ã®ã‚¤ãƒ»ã‚»ãƒ‰ãƒ«ï¼™æ®µã¨ã®ï¼•ç•ªå‹è² ã§ï¼“é€£å‹ã—ã€å‹ã¡è¶Šã—ã¾ã—ãŸã€‚ã‚¤ï¼™æ®µã¯æ®‹ã‚‹ï¼’å±€ã§ä¸€çŸ¢å ±ã„ãŸã„ã¨ã—ã¦ã„ã¾ã™ã€‚ã“ã®æ—¥ã¯é–“é•ã„ãªãæ´å²ã«åˆ»ã¾ã‚Œã‚‹ã§ã—ã‚‡ã†ã€‚ ã€Œãƒ‡ã‚£ãƒ¼ãƒ—ãƒžã‚¤ãƒ³ãƒ‰ã€ã®å…±åŒå‰µæ¥è€…ãƒ‡ãƒŸã‚¹ãƒ»ãƒã‚µãƒ“ã‚¹æ°ã¯ã€Œæ´å²çš„ãªçž¬é–“ã ã€ã¨ãƒ„ã‚£ãƒ¼ãƒˆã—ã¦ã„ã¾ã™ã€‚ãªãœã€Œæ´å²çš„ãªçž¬é–“ã€ãªã®ã‹ã€äººå·¥çŸ¥èƒ½ã®æ´å²ã‚’æŒ¯ã‚Šè¿”ã‚Šã¾ã—ã‚‡ã†ã€‚ 1950å¹´ã€€ãƒŠãƒã‚¹ãƒ‰ã‚¤ãƒ„ã®æš—å·æ©Ÿã‚¨ãƒ‹ã‚°ãƒžã®è§£èªã«ã‚‚æˆåŠŸã—ãŸè‹±å›½ã®å¤©æ‰æ•°å¦è€…ã‚¢ãƒ©ãƒ³ãƒ»ãƒãƒ¥ãƒ¼ãƒªãƒ³ã‚°æ°ãŒã€Œãƒãƒ¥ãƒ¼ãƒªãƒ³ã‚°ãƒ»ãƒ†ã‚¹ãƒˆã€ã‚’ææ¡ˆã€‚äººé–“ãŒéŸ³å£°ã‚’åŒã˜ã«ã—ãŸäººé–“ã¨æ©Ÿæ¢°ã¨ãã‚Œãžã‚Œå¯¾è©±ã—ã¦åŒºåˆ¥ã§ããªã‘ã‚Œã°ã€Œäººå·¥çŸ¥èƒ½ã€ã¨èªã‚ã‚‰ã‚Œã‚‹ã¨ã„ã†åŸºæº–ã‚’ç¤ºã—ãŸ 1952å¹´ã€€ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ãƒ¼ãŒãƒžãƒ«ãƒãƒ„ã‚²ãƒ¼ãƒ ã‚’ç¿’å¾—ã™ã‚‹ 1994å¹´ã€€ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ãƒ¼ãŒãƒœãƒ¼ãƒ‰ï¼ˆç›¤ä¸Šï¼‰ã‚²ãƒ¼ãƒ ã®ãƒã‚§
nfunato 2016/03/13
igo

deeplearning
ãƒªãƒ³ã‚¯
ã€ŽAlphaGo VS ã‚¤ã‚»ãƒ‰ãƒ«ï¼ˆç¬¬2æˆ¦ï¼‰æ„Ÿæƒ³ ã€€ç¥žã¯æ»ã‚“ã ã€‚ã€
æ‘ä¸Šæ·±ã®æ°—ãŒå‘ã„ãŸã¨ãã«æ›´æ–°ã™ã‚‹ãƒ–ãƒã‚°ï¼šæœˆï¼’ç¢ï¼ˆè„±ã‚µãƒ©â†’å›²ç¢ç•Œã®ä½äººï¼‰ ï¼—å¹´é–“å‹¤ã‚ãŸå¯Œå£«é€šã‚’ï¼’ï¼ï¼‘ï¼–å¹´ï¼‘æœˆã«é€€è·ã—ã€å›²ç¢ã®ä¸–ç•Œã‚’ä¿ºãŒå¤‰ãˆã‚‹ã‚“ã ï¼ã¨ã„ã†ãƒŽãƒªã§ä»Šæ—¥ã‚‚ãƒžã‚¤ãƒšãƒ¼ã‚¹ã«ç”Ÿãã¦ã¾ã™ã€‚ ï¼’ï¼æ³é ƒã¾ã§å›²ç¢æ£‹å£«ã‚’ç›®æŒ‡ã—ã¦ã¾ã—ãŸã€‚ ï¼Šã“ã®è¨˜äº‹ã¯æ£‹èœãŠã‚ˆã³ä¸‰æ‘ä¹æ®µã®è§£èª¬è¨˜äº‹ã€ã€Œå¹½çŽ„ã®é–“ã€ã®çŽ‹ãƒ¡ã‚¤ã‚¨ãƒ³ä¹æ®µã®è§£èª¬ã‚’å‚è€ƒã«è¨˜è¿°ã—ã¦ãŠã‚Šã¾ã™ã€‚ä»–ã®ãƒ—ãƒæ£‹å£«ã®æ–¹ã®è§£èª¬ãªã©ã¯å‚ç…§ã—ã¦ãŠã‚Šã¾ã›ã‚“ã®ã§ã€äºˆã‚ã”äº†æ‰¿ãã ã•ã„ã€‚ ã©ã‚‚ã§ã™ã€‚ 2å±€ç›®ã‚‚Alpha GoãŒå‹ã¡ã¾ã—ãŸãã€‚ æ£ç›´ã€æ€ã„ã®ã»ã‹ã€æ„Ÿå‹•ã—ã¾ã—ãŸã€‚ ä¿ºã¯ç„¡ç¥žè«–è€…ã§ã™ã€‚ç„¡ç¥žè«–ã¨ã„ã†ã®ã¯ã€ç‰¹å®šã®å®—æ•™è¦³ã‚’æŒãŸãªã„ã¨ã„ã†æ„å‘³ã§ã¯ãªãã€ã€Œç¥žã¯ã„ãªã„ã€ã¨ã„ã†å®—æ•™è¦³ã¨ã„ã†ã“ã¨ã§ã™ãŒã€‚ å›²ç¢ã¯å®‡å®™ã«ä¾‹ãˆã‚‰ã‚Œã‚‹ã“ã¨ãŒå¤šã„ã§ã™ã€‚ å¤‰åŒ–ãŒéžå¸¸ã«å¤§ããã€äººé–“ã§ã¯çŸ¥è¦šã—ãã‚Œãªã„ã»ã©ã®æƒ…å ±é‡ã‚’ç§˜ã‚ã¦ã„ã‚‹ã‹ã‚‰ã€ã¨ã„ã†ã®ãŒè«–æ‹ ã§ã‚ã‚‹å ´åˆãŒå¤šã„ã®ã§ã™ãŒã€å˜ç´”ã«ãƒãƒžãƒ³ã‚’è¨€ã„è¡¨ã—ã¦ã„ã‚‹ã“ã¨ã‚‚
nfunato 2016/03/11
igo
ãƒªãƒ³ã‚¯
AlphaGoã¨ãƒˆãƒƒãƒ—æ£‹å£«ã®å¯¾å±€ã‚’è¦³æˆ¦ã—ã‚ˆã†ï¼ - ã‚ã‚“ã¡ã¹ï¼
"We cannot go against the trend. I think machines will beat humans someday. " "If I get defeated it might be negative for go and this is inevitable in this modern life. However, it will not destroy the value of go itself." ï¼ˆã€Œç§ãŸã¡ã¯æ™‚ä»£ã®æµã‚Œã«é€†ã‚‰ã†ã“ã¨ã¯å‡ºæ¥ã¾ã›ã‚“ã€‚ã„ã¤ã‹æ©Ÿæ¢°ã¯äººé–“ã‚’æ‰“ã¡è² ã‹ã™ã§ã—ã‚‡ã†ã€ ã€Œã‚‚ã—ç§ãŒè² ã‘ãŸã‚‰å›²ç¢ã«ã¨ã£ã¦ã¯ãƒã‚¬ãƒ†ã‚£ãƒ–ãªã“ã¨ã§ã™ãŒã€ã“ã‚Œã¯ç¾ä»£ç¤¾ä¼šã§ã¯é¿ã‘ã‚‰ã‚Œãªã„ã“ã¨ã§ã™ã€‚ ã—ã‹ã—ã€ãã®æ•—åŒ—ã¯å›²ç¢è‡ªèº«ã®ä¾¡å€¤ã‚’æãªã†ã“ã¨ã¯ãªã„ã§ã—ã‚‡ã†ã€ã‚¤ãƒ»ã‚»ãƒ‰ãƒ«ï¼‰ Human-AI showdown begins 1 p.m. 2016å¹´3æœˆ9æ—¥ã€"Alph
nfunato 2016/03/10
igo
ãƒªãƒ³ã‚¯
1 2 æ¬¡ã®ãƒšãƒ¼ã‚¸