[B! ACL] jewel12ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

jewel12 id:jewel12

ACLã«é–¢ã™ã‚‹jewel12ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (4)

${{author_name}}$

{{{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}}

ACL 2011 ã§å…¬é–‹ã•ã‚ŒãŸè‡ªç„¶è¨€èªžå‡¦ç†ï¼ˆè¨€ã„æ›ãˆç³»ï¼‰ãƒ‡ãƒ¼ã‚¿
å…ˆé€±ã‚ªãƒ¬ã‚´ãƒ³å·žãƒãƒ¼ãƒˆãƒ©ãƒ³ãƒ‰ã§è¡Œã‚ã‚ŒãŸACL2011ã¨ã„ã†å›½éš›ä¼šè°ã§ç™ºè¡¨ã•ã‚ŒãŸãƒ‡ãƒ¼ã‚¿ã®ã†ã¡ã€ã™ãã«ãƒ€ã‚¦ãƒ³ãƒãƒ¼ãƒ‰ã—ã¦é¢ç™½ã„å®Ÿé¨“ãŒã§ããã†ãªè¨€ã„æ›ãˆç³»ã®ãƒ‡ãƒ¼ã‚¿ã‚’ï¼•ã¤ã‚’ç´¹ä»‹ã—ã¾ã™ã€‚ï¼ˆä»–ã«ã‚‚ã„ã‚ã„ã‚é¢ç™½ã„ç™ºè¡¨ã¯ã‚ã£ãŸã®ã§ã™ãŒã€ãƒ†ãƒ¼ãƒžã‚’çµžã£ã¦ã¿ã¾ã—ãŸã€‚ã“ã®ãƒ–ãƒã‚°ã®èªè€…ã«ã©ã‚Œã ã‘éœ€è¦ã‚ã‚‹ã‹ã‚ã‹ã‚Šã¾ã›ã‚“ãŒãƒ»ãƒ»ãƒ»ã€‚ï¼‰ ã€1ã€‘ Session 1-E -Â Collecting Highly Parallel Data for Paraphrase Evaluation David Chen1Â and William Dolan2 1The University of Texas at Austin,Â 2Microsoft Research è¤‡æ•°ã®ãƒ¡ã‚«ãƒ‹ã‚«ãƒ«ã‚¿ãƒ¼ã‚«ãƒ¼ãŒåŒã˜å‹•ç”»ã«ä»˜ã‘ãŸæ³¨é‡ˆã‚’ãƒ‘ãƒ©ãƒ¬ãƒ«ã‚³ãƒ¼ãƒ‘ã‚¹ã«ã—ãŸã‚‚ã®ã€‚ï¼ˆå‹•ç”»ã ã¨ç”»åƒã‚ˆã‚Šè§£é‡ˆã®ãƒ–ãƒ¬ãŒæ¸›ã‚‹ã¨ã‹ã€‚ï¼‰è¨€ã„æ›ãˆã®ã¿ãªã‚‰ãšåŒã˜äº‹æŸ„ã®é•ã†è§£é‡ˆã‚’å«ã‚€ãŸã‚ã€ã±ã£ã¨
jewel12 2011/07/03
acl

data

nlp
ãƒªãƒ³ã‚¯
ACL HLT 2011 2æ—¥ç›®: NLPã®ãƒˆãƒ¬ãƒ³ãƒ‰ã¯ã‚°ãƒ©ãƒ•ã‚’ç”¨ã„ãŸæ‰‹æ³• - æ¦è”µé‡Žæ—¥è¨˜
5æ™‚ã«æ—¥å…‰ã§è‡ªç„¶ã«ç›®ãŒè¦šã‚ãŸã®ã§é›»è©±ã€‚å‰ä½¿ã£ã¦ã„ãŸãƒ—ãƒªãƒšã‚¤ãƒ‰ã®æºå¸¯ã‚’æŒã£ã¦ããŸã®ã§$30ãƒãƒ£ãƒ¼ã‚¸ã—ã¦ã€ã‹ã‚Œã“ã‚Œ3æ™‚é–“ãã‚‰ã„è©±ã—ã¦ã„ã‚‹ã¨æ€ã†ã®ã ãŒã€ã¾ã ãªããªã‚‰ãªã„ã€‚æ—¥æœ¬ã®æºå¸¯é›»è©±ã«ã‹ã‘ã¦1åˆ†$0.15ã®ãƒ¬ãƒ¼ãƒˆãªã®ã§ã€200åˆ†ã„ã‘ã‚‹ã‚ˆã†ã ã€‚ æœã¯ãƒ™ã‚¹ãƒˆãƒšãƒ¼ãƒ‘ãƒ¼ãƒ»ãƒšã‚¹ãƒˆã‚¹ãƒãƒ¥ãƒ¼ãƒ‡ãƒ³ãƒˆãƒšãƒ¼ãƒ‘ãƒ¼ãƒ»ãƒ™ã‚¹ãƒˆã‚·ãƒ§ãƒ¼ãƒˆãƒšãƒ¼ãƒ‘ãƒ¼ã®3æœ¬ã®ãƒˆãƒ¼ã‚¯ã€‚ã“ã‚Œã‚‰ã®è«–æ–‡ãŒæ±ºã¾ã‚‰ãªã„ã¨ä¼šè°ã®ãƒ—ãƒã‚°ãƒ©ãƒ ã‚’ç¢ºå®šã§ããªã„ã®ã§ã€æ—©ããƒ—ãƒã‚°ãƒ©ãƒ ã‚’å…¬é–‹ã—ã¦ã»ã—ã„ã€ã¨ã„ã‚ã‚“ãªäººã‹ã‚‰ã›ã£ã¤ã‹ã‚Œã¦å›°ã£ãŸã€ã¨ (ACL ã®ãƒ—ãƒã‚°ãƒ©ãƒ å§”å“¡é•·) æ¾æœ¬å…ˆç”ŸãŒã¼ã‚„ã„ã¦ã„ãŸã®ã ãŒã€ãƒ™ã‚¹ãƒˆãƒšãƒ¼ãƒ‘ãƒ¼ã®ç´¹ä»‹ã‚’ä»Šæ—¥ã¯æ¾æœ¬å…ˆç”ŸãŒã—ã¦ã€ãƒ™ã‚¹ãƒˆãƒšãƒ¼ãƒ‘ãƒ¼ã‚’äº‹å‰ã«ç™ºè¡¨ã™ã‚‹ Pros and Cons ã¨ã„ã†ã‚¿ã‚¤ãƒˆãƒ«ã§ã—ã£ã‹ã‚Šã€Œäº‹å‰ã«æ±ºã‚ã‚‹ã®ã¯å¤§å¤‰ã ã£ãŸã€ã¨è‹¦è¨€ã‚’å‘ˆã—ã¦ã„ãŸ (ç¬‘)ã€€å‰¯å¦é•·ã«åŠ ãˆã¦ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°å§”å“¡é•·ã€ãªã‚“ã ã‹ä»Šå¹´æ¾æœ¬å…ˆç”Ÿã¯å¿™ã—ãã†ãªå¹´ã§ã‚ã‚‹ã€‚ ã•ã¦ãƒ™ã‚¹ãƒˆ
jewel12 2011/06/27
ACL
ãƒªãƒ³ã‚¯
ACL HLT 2011 åˆæ—¥: äººé–“ã®ã‚¯ã‚¤ã‚ºçŽ‹ã«å‹ã£ãŸ IBM Watson ã‚’æ”¯ãˆã‚‹æŠ€è¡“ - æ¦è”µé‡Žæ—¥è¨˜
ä»Šæ—¥ã¯è‡ªç„¶è¨€èªžå‡¦ç†ã«é–¢ã™ã‚‹ä¸–ç•Œæœ€å¤§ã®ä¼šè°ã€ACL ã®æœ¬ä¼šè°åˆæ—¥ã€‚ãƒ›ãƒ†ãƒ«é–‹å‚¬ãªã®ã§æœç°¡å˜ã«ã¤ã¾ã‚ã‚‹ã‚‚ã®ã¯ä¼šå ´ã«ã‚ã‚‹ã®ã§ã¯ã€ã¨æ€ã£ã¦è¡Œã£ãŸã‚‰æ¡ˆã®å®šãƒ™ãƒ¼ã‚°ãƒ«ã¨ã‹ã‚³ãƒ¼ãƒ’ãƒ¼ã¨ã‹ãƒ•ãƒ«ãƒ¼ãƒ„ãŒã‚ã£ãŸã®ã§ã€ãƒ™ãƒ¼ã‚°ãƒ«ã‚’é£Ÿã¹ã‚‹ã€‚ã‘ã£ã“ã†ã„ã‘ã‚‹ã€‚ ã‚ªãƒ¼ãƒ—ãƒ‹ãƒ³ã‚°ã§ã¯ä»Šå›žã®ä¼šè°ã®çµ±è¨ˆã‚’è©±ã—ã¦ã„ãŸãŒã€1,146æœ¬ã®æŠ•ç¨¿(ã†ã¡646æœ¬ãŒãƒ•ãƒ«ãƒšãƒ¼ãƒ‘ãƒ¼ã€æ®‹ã‚Šã®512æœ¬ãŒã‚·ãƒ§ãƒ¼ãƒˆãƒšãƒ¼ãƒ‘ãƒ¼)ã€292æœ¬(ã†ã¡164æœ¬ãŒãƒ•ãƒ«ãƒšãƒ¼ãƒ‘ãƒ¼ã€128æœ¬ãŒã‚·ãƒ§ãƒ¼ãƒˆãƒšãƒ¼ãƒ‘ãƒ¼)ãŒæŽ¡æŠžã¨ã®ã“ã¨ã€‚ãƒ•ãƒ«ãƒšãƒ¼ãƒ‘ãƒ¼(2æ®µçµ„8ãƒšãƒ¼ã‚¸+å‚è€ƒæ–‡çŒ®)ã¨ã‚·ãƒ§ãƒ¼ãƒˆãƒšãƒ¼ãƒ‘ãƒ¼(2æ®µçµ„4ãƒšãƒ¼ã‚¸+å‚è€ƒæ–‡çŒ®)ã¨ã‚‚ã«æŽ¡æŠžçŽ‡ã¯25%ã€‚è‡ªç„¶è¨€èªžå‡¦ç†ã§ã¯æŽ¡æŠžçŽ‡ãŒ40%ãã‚‰ã„ã®å›½éš›ä¼šè°ãŒå¤šã„ã®ã§ã€25%ã¯ã‚„ã¯ã‚Šé›£é–¢ä¼šè°ã§ã‚ã‚‹ã€‚ æ‹›å¾…è¬›æ¼”ã¯ã‚¢ãƒ¡ãƒªã‚«ã®ã‚¯ã‚¤ã‚ºç•ªçµ„ã§äººé–“ã«å‹ã£ãŸã“ã¨ã§æœ‰åã«ãªã£ãŸIBM Watson ã‚’æ”¯ãˆã‚‹æŠ€è¡“ã«ã¤ã„ã¦ã€‚ãƒã‚¿ãŒã‚ã¾ã‚Šã‚ˆãåˆ†ã‹ã‚‰ãšã„ã¾ã„ã¡æ¥½ã—ã‚ãªã‹ã£ãŸãŒ (ç´°ã‹ã„
jewel12 2011/06/22
ACL

nlp
ãƒªãƒ³ã‚¯
Welcome to ACL 2010 Conference
jewel12 2010/04/27
nlp

acl
ãƒªãƒ³ã‚¯
1