[B! algorithm] pcodã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

pcod id:pcod

algorithmã«é–¢ã™ã‚‹pcodã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (197)

${{author_name}}$

{{{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}}

ä¸å¦ç”Ÿã«ã‚‚ã‚ã‹ã‚‹ã‚¦ã‚§ãƒ¼ãƒ–ãƒ¬ãƒƒãƒˆè¡Œåˆ— - ã‚¢ã‚¹ãƒšæ—¥è¨˜
id:echizen_tm ã•ã‚“ã®è¨˜äº‹ã€Œã‚¦ã‚§ãƒ¼ãƒ–ãƒ¬ãƒƒãƒˆæœ¨ã®åŠ¹çŽ‡çš„ã§ç°¡å˜ãªå®Ÿè£… "The Wavelet Matrix"ã€ã‹ã‚‰å§‹ã¾ã£ãŸã‚¦ã‚§ãƒ¼ãƒ–ãƒ¬ãƒƒãƒˆè¡Œåˆ—ãƒ–ãƒ¼ãƒ ã‹ã‚‰åŠå¹´ä»¥ä¸ŠãŒéŽãŽã€ã™ã§ã«æž¯ã‚ŒãŸæŠ€è¡“ã¨ã—ã¦ç¢ºç«‹ã•ã‚Œã¤ã¤ã‚ã‚‹æ„ŸãŒã‚ã‚Šã¾ã™ã€‚ â€¦å˜˜ã§ã™ã€‚ æ—¥æœ¬ä»¥å¤–ã§ã¯ã‚ã‚“ã¾ã‚Šæ¥ã¦ã„ã¾ã›ã‚“ã€‚ ç†ç”±ã¨ã—ã¦ã¯ã€ã‚„ã¯ã‚Šã‚¢ãƒ«ãƒ•ã‚¡ãƒ™ãƒƒãƒˆåœã§ã¯å˜èªžå¢ƒç•ŒãŒæ˜Žç¢ºã§ã‚ã‚‹ãŸã‚ã€ã“ã¡ã‚‰ã®è¨˜äº‹ã§æ›¸ã‹ã‚Œã¦ã„ã‚‹ã‚ˆã†ãªã€Œã‚ãƒ¼ãƒ¯ãƒ¼ãƒ‰åˆ†å‰²ã®é›£æ˜“åº¦ã€ã¨ã„ã£ãŸã“ã¨ãŒã‚ã¾ã‚Šå•é¡Œã«ãªã‚‰ãªã„ã¨ã„ã†ã“ã¨ãŒã‚ã‚‹ã‹ã‚‚ã—ã‚Œã¾ã›ã‚“ã€‚ ã¾ã‚ã€ãã†ã„ã†ã‚ã‘ã§å±€æ‰€çš„ã«æ¥ã¦ã„ã‚‹ã‚¦ã‚§ãƒ¼ãƒ–ãƒ¬ãƒƒãƒˆè¡Œåˆ—ã§ã™ãŒã€æ—¥æœ¬èªžã‚’ã¯ã˜ã‚ã¨ã™ã‚‹å˜èªžå¢ƒç•Œã®ãªã„è¨€èªžåœã«ã¨ã£ã¦ã¯é‡è¦ãªãƒã‚¿ã§ã‚ã‚‹ã¨æ€ã†ãŸã‚ã€è§£èª¬è¨˜äº‹ã‚’æ›¸ãç›´ã—ã¦*1ã¿ã‚ˆã†ã¨æ€ã„ã¾ã™ã€‚ ã‚¦ã‚§ãƒ¼ãƒ–ãƒ¬ãƒƒãƒˆè¡Œåˆ—ã§ã§ãã‚‹ã“ã¨ ä¸»ã¨ãªã‚‹æ“ä½œã¯ã€æ–‡å—åˆ—ã«å¯¾ã™ã‚‹ å®šæ•°æ™‚é–“ã® rank() ã¨ select()*2 ã§ã™ã€‚ rank() ã¯ã€ã€Œæ–‡
pcod 2013/11/20
algorithm
ãƒªãƒ³ã‚¯
Algorithm::LibLinear ã®ç´¹ä»‹
Notice: This article is outdated. Please refer an updated English tutorial. è¦æ—¨ã‹ãªã‚Šå‰ã«ãªã‚Šã¾ã™ãŒã€Algorithm::LibLinear ã¨ã„ã† Perl ãƒ¢ã‚¸ãƒ¥ãƒ¼ãƒ«ã‚’æ›¸ãã¾ã—ãŸã€‚ CPAN Github ã“ã‚Œã‚’ä½¿ã†ã¨ç·šå½¢åˆ†é¡žå™¨ãªã©ãŒé«˜é€Ÿã«å¦ç¿’ã§ãã¾ã™ã€‚ãƒ†ã‚ã‚¹ãƒˆã‚„ç”»åƒã®åˆ†é¡žãŒå¿œç”¨ã¨ã—ã¦æœŸå¾…ã•ã‚Œã¾ã™ã€‚ LIBLINEAR ã«ã¤ã„ã¦LIBLINEAR ã¯ LIBSVM ã¨åŒã˜å°æ¹¾å›½ç«‹å¤§å¦ã® Chih-Jen Lin æ•™æŽˆã®ãƒãƒ¼ãƒ ãŒå…¬é–‹ã—ã¦ã„ã‚‹ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹ã®æ©Ÿæ¢°å¦ç¿’ãƒ‘ãƒƒã‚±ãƒ¼ã‚¸ã§ã™ã€‚ é–¢æ•°ã®ãƒã‚¸ã‚¹ãƒ†ã‚£ãƒƒã‚¯å›žå¸°ã€ã‚µãƒãƒ¼ãƒˆãƒ™ã‚¯ã‚¿ãƒ¼å›žå¸°åŠã³ç·šå½¢ SVM ã«ã‚ˆã‚‹å¤šã‚¯ãƒ©ã‚¹åˆ†é¡žã‚’è¡Œã†ã“ã¨ãŒã§ãã¾ã™ã€‚LIBSVM ã¨é•ã£ã¦ã‚«ãƒ¼ãƒãƒ«é–¢æ•°ã‚’ä½¿ã†ã“ã¨ã¯ã§ãã¾ã›ã‚“ãŒã€ã¯ã‚‹ã‹ã«é«˜é€Ÿã«å‹•ä½œã—ã¾ã™ã€‚ Algorithm::LibLinear
pcod 2013/08/13
algorithm
ãƒªãƒ³ã‚¯
b-Bit MinHashã«ã‚ˆã‚‹é«˜é€Ÿã‹ã¤çœã‚¹ãƒšãƒ¼ã‚¹ãªé¡žä¼¼åº¦åˆ¤å®š | SmartNewsé–‹ç™ºè€…ãƒ–ãƒã‚°
ã‚´ã‚¯ãƒã®æµœæœ¬ã§ã™ã€‚ãƒãƒƒãƒˆã‚«ãƒ•ã‚§ã§ã‚³ãƒ¼ãƒ‰ã‚’æ›¸ãã®ãŒå¥½ãã§ã™ã€‚ å‰å›žã®ã‚¨ãƒ³ãƒˆãƒªãƒ¼ã§ã‚‚è§¦ã‚Œã‚‰ã‚Œã¦ã„ã¾ã™ãŒã€SmartNewsã¯ãƒ›ãƒƒãƒˆãªè©±é¡Œã‚’ãƒ¦ãƒ¼ã‚¶ã«ãŠå±Šã‘ã™ã‚‹ãŸã‚ã«ã€å¸¸æ™‚ã€è†¨å¤§ãªæ•°ã®ãƒ„ã‚¤ãƒ¼ãƒˆãŠã‚ˆã³URLã‚’ã‚¯ãƒãƒ¼ãƒ«ã—ã¦ã„ã¾ã™ã€‚ã“ã†ã—ã¦åŽé›†ã—ãŸè¨˜äº‹ã«å¯¾ã—ã€æ§˜ã€…ãªåˆ†æžãŒæ–½ã•ã‚Œã¾ã™ãŒã€ãã®ä¸ã§ã‚‚é‡è¦ãªå‡¦ç†ã®1ã¤ã«ã€è¨˜äº‹ã®é¡žä¼¼åº¦åˆ¤å®šãŒã‚ã‚Šã¾ã™ã€‚å†…å®¹ã®ä¼¼é€šã£ãŸè¨˜äº‹ã‚’ã‚¤ãƒ³ãƒ‡ãƒƒã‚¯ã‚¹ã‹ã‚‰ç™ºè¦‹ã—ã€ã‚°ãƒ«ãƒ¼ãƒ—åŒ–ã™ã‚‹å‡¦ç†ã§ã™ã€‚ æ¯Žç§’ã€å¤§é‡ã®æ–°ç€è¨˜äº‹ãŒåˆ°ç€ã™ã‚‹ã“ã¨ã‹ã‚‰ã€ã“ã®é¡žä¼¼åº¦åˆ¤å®šã¯é«˜é€Ÿã«å®Ÿè¡Œã™ã‚‹å¿…è¦ãŒã‚ã‚Šã¾ã™ã€‚ã¾ãŸã€ã‚¤ãƒ³ãƒ‡ãƒƒã‚¯ã‚¹ã‚’å…¨ã¦ãƒ¡ãƒ¢ãƒªã«è¼‰ã›ã¦ã„ã‚‹ã®ã§ã€é¡žä¼¼åº¦åˆ¤å®šã‚’å®Ÿç¾ã™ã‚‹éš›ã®ç©ºé–“åŠ¹çŽ‡ã‚‚è¦æ±‚ã•ã‚Œã¾ã™ã€‚ ä»Šå›žã¯ã€SmartNewsãŒé«˜é€Ÿã‹ã¤çœã‚¹ãƒšãƒ¼ã‚¹ãªé¡žä¼¼åº¦åˆ¤å®šã®ãŸã‚ã«ä½¿ç”¨ã—ã¦ã„ã‚‹b-Bit MinHashã¨å‘¼ã°ã‚Œã‚‹æ‰‹æ³•ã‚’ç´¹ä»‹ã—ã¾ã™ã€‚2å¹´å‰ã«ã€PFIã®å²¡é‡ŽåŽŸã•ã‚“ãŒéžå¸¸ã«åˆ†ã‹ã‚Šã‚„ã™ã„è§£èª¬è¨˜äº‹ã‚’æ›¸ã‹ã‚Œã¦ãŠã‚Šã€æœ¬ã‚¨ãƒ³ãƒˆãƒªãƒ¼
pcod 2013/08/05
algorithm
ãƒªãƒ³ã‚¯
ã“ã®ãƒšãƒ¼ã‚¸ã‚’è¦‹ã‚‹ã«ã¯ã€ãƒã‚°ã‚¤ãƒ³ã¾ãŸã¯ç™»éŒ²ã—ã¦ãã ã•ã„
Facebookã§æŠ•ç¨¿ã‚„å†™çœŸãªã©ã‚’ãƒã‚§ãƒƒã‚¯ã§ãã¾ã™ã€‚
pcod 2013/04/15
algorithm

facebook

search
ãƒªãƒ³ã‚¯
ã€ŽMaxentSets ã‚’ä½¿ã£ãŸèŠ¸èƒ½äººæŽ¨è–¦ã€€â€•ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ ã‚·ãƒªãƒ¼ã‚º 2â€•ã€
åº¦ã€…ç™»å ´ã•ã›ã¦é ‚ã„ã¦ãŠã‚Šã¾ã™ hattori ã§ã™ã€‚ ãƒœã‚¹ã‹ã‚‰ã®ãŠé”ã—ã§2å›žç›®ã®ãƒã‚¿ã‚’æ›¸ãã“ã¨ã«ãªã‚Šã¾ã—ãŸã€‚( å‰å›žã®ãƒã‚¿ã¯ã“ã¡ã‚‰ ) èžãæ‰€ã«ã‚ˆã‚‹ã¨3ãƒ¶æœˆã«ä¸€å›žãã‚‰ã„ã®ãƒšãƒ¼ã‚¹ã§æœ¬ã‚·ãƒªãƒ¼ã‚ºï¼Ÿã‚’æ›¸ã‹ãã°ãªã‚‰ãªã„ã®ã§ã™ãŒã€1å¹´å¾Œã«æžœãŸã—ã¦æ›¸ããƒã‚¿ãŒæ®‹ã£ã¦ã„ã‚‹ã‹ã€ã€ã€è‹¥å¹²å¾®å¦™ãªæ„Ÿã˜ã§ã™ã€‚æ±—ã€‚ ã•ã¦æœ¬æ—¥ã®ãƒ†ãƒ¼ãƒžã§ã™ãŒã€" Set Expansionã‚’åˆ©ç”¨ã—ãŸèŠ¸èƒ½äººæŽ¨è–¦ã®æ–¹æ³• " ã‚’ç´¹ä»‹ã—ã‚ˆã†ã¨æ€ã„ã¾ã™ã€‚ã“ã‚Œã¯ä»¥å‰ã€æ ªå¼ä¼šç¤¾ mixi æ§˜ã¨ã®åˆåŒå‹‰å¼·ä¼šã§ç™ºè¡¨ã—ãŸå†…å®¹ã®è£œè¶³çš„ãªè©±ã«ãªã£ã¦ã„ã¾ã—ã¦ã€ä»¥ä¸‹ã®ç™ºè¡¨ã‚¹ãƒ©ã‚¤ãƒ‰ã‚’å…ˆã«æ–œã‚èªã¿ã—ã¦é ‚ã‘ã¾ã™ã¨ä»¥é™ã®è©±ãŒåˆ†ã‹ã‚Šæ˜“ã„ã‹ã¨æ€ã„ã¾ã™ã€‚ â– ã€€Set Expansion ã¨ã¯ä½•ãžã‚„ï¼Ÿ Set Expansion ã¨ã¯ ã€Ž SEEDã¨ãªã‚‹éƒ¨åˆ†ã‚»ãƒƒãƒˆã‚’æ‹¡å¼µãƒ»è£œå®Œã—ã¦ã€ã‚ˆã‚Šå®Œå…¨ãªã‚»ãƒƒãƒˆé›†åˆã‚’ä½œã‚‹ ã€ ã¨ã„ã†ã‚¿ã‚¹ã‚¯ã®äº‹ã§ã™ã€‚è‹¥å¹²åˆ†ã‹ã‚Šã«é›£ã„ã®ã§ã™ãŒã€ã‚ˆã‚Šå¹³æ˜“ã«è¨€ã†ã¨ ã€Ž
pcod 2013/03/06
ã‚¹ãƒ—ãƒ¬ãƒƒãƒ‰ã‚·ãƒ¼ãƒˆã®è‡ªå‹•è£œå®Œã¿ã¦SEALæ€ã„å‡ºã—ãŸãŒã€ã‚‚ã£ã¨é¢ç™½ã„ã‚‚ã®ãŒã‚ã£ãŸ

algorithm
ãƒªãƒ³ã‚¯
Efï¬cient Personalized PageRank with Accuracy Assurance
pcod 2013/02/15
algorithm
ãƒªãƒ³ã‚¯
è¨ˆé‡å¦ç¿’ã‚’ç”¨ã„ãŸç”»åƒæ¤œç´¢ã‚¨ãƒ³ã‚¸ãƒ³ã¨ã‚¢ãƒ‹ãƒ¡é¡”é¡žä¼¼æ¤œç´¢v3ã«ã¤ã„ã¦ - ãƒ‡ãƒ¼
ã¾ã githubã«ã¯pushã—ã¦ã„ãªã„ã®ã§ã™ãŒã€ã•ã„ãã‚‡ã†ã®çµ„ã¿è¾¼ã¿åž‹ç”»åƒæ¤œç´¢ã‚¨ãƒ³ã‚¸ãƒ³otamaã«è¨ˆé‡å¦ç¿’ã‚’ç”¨ã„ã¦ä¸Žãˆã‚‰ã‚ŒãŸãƒ‡ãƒ¼ã‚¿ã«ã‚ã£ãŸç”»åƒé–“ã®è·é›¢é–¢æ•°ã‚’å¦ç¿’ã—ã¦ãã‚Œã‚’ä½¿ã£ã¦æ¤œç´¢ã™ã‚‹ã¨ã„ã†ãƒ‰ãƒ©ã‚¤ãƒã‚’å…¥ã‚ŒãŸã®ã§ã€å…ˆè¡Œçš„ãªãƒ‡ãƒ¢ã¨ã—ã¦ã‚¢ãƒ‹ãƒ¡é¡”é¡žä¼¼æ¤œç´¢v3ã‚’ä½œã£ã¦ã¿ã¾ã—ãŸã€‚ è¨ˆé‡å¦ç¿’ã¯ã€ãƒ™ã‚¯ãƒˆãƒ«é–“ã®è·é›¢ã®è¨ˆã‚Šæ–¹ã‚’æ©Ÿæ¢°å¦ç¿’ã§æ±ºã‚ã‚‹ã¿ãŸã„ãªåˆ†é‡Žã§ã™ã€‚ ã‚¢ãƒ‹ãƒ¡é¡”é¡žä¼¼æ¤œç´¢v3 AnimeFace Search v3 - Otama LMCA_VLAD_HSV Driver randomãƒœã‚¿ãƒ³ã‚’æŠ¼ã™ã¨é¡”ç”»åƒãŒãƒ©ãƒ³ãƒ€ãƒ ã«å‡ºã‚‹ã®ã§ã©ã‚Œã‹ã‚¯ãƒªãƒƒã‚¯ã™ã‚‹ã¨ãã‚Œã‚’ã‚¯ã‚¨ãƒªã«æ¤œç´¢ã—ã¾ã™ã€‚color weightã¯è‰²ã®é‡ã¿ã‚’èª¿ç¯€ã™ã‚‹ãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿ãƒ¼ã§ã€1ã«ã™ã‚‹ã¨è‰²ã ã‘ã§æ¤œç´¢ã—ã¾ã™ã€‚0ã«ã™ã‚‹ã¨å½¢çŠ¶ã‚„ãƒ†ã‚¯ã‚¹ãƒãƒ£ã ã‘ã§æ¤œç´¢ã—ã¾ã™ã€‚çµæžœç”»åƒã®ä¸Šã®æ•°å—ã¯é¡žä¼¼åº¦çš„ãªã‚‚ã®ã§ã€ãã®æ¨ªã®gglã¯å…ƒç”»åƒã‚’Google Search by Imag
pcod 2013/01/10
search

algorithm

ç”»åƒå‡¦ç†
ãƒªãƒ³ã‚¯
PNGè»½é‡åŒ–ã®æ¸›è‰²ã¨åœ§ç¸®ã«ã¤ã„ã¦ | GREE Engineering
ã“ã®ãƒ†ãƒ¼ãƒ–ãƒ«ã®ç•ªå·ã¯ 1 Byte ã«ãªã£ã¦ã„ã‚‹ãŸã‚ã€0-255 ã® 256 å€‹ã—ã‹ç™»éŒ²ã§ãã¾ã›ã‚“ã€‚ãã®ãŸã‚ã€ç”»åƒã§ä½¿ç”¨ã•ã‚Œã¦ã„ã‚‹è‰²ãŒ 256 å€‹ã‚ˆã‚Šå¤šã„å ´åˆã¯ã€ãªã‚“ã¨ã‹ã—ã¦ 256 å€‹ã«ã—ãªãã¦ã¯ã„ã‘ã¾ã›ã‚“ã€‚ ã“ã®ã€Œãªã‚“ã¨ã‹ã—ã¦ 256 è‰²ã«ã™ã‚‹ã€ã¨ã„ã†ã®ãŒæ¸›è‰²å‡¦ç†ã§ã€ãªã‚‹ã¹ãå…ƒã®ç”»åƒã‹ã‚‰ã®å¤‰åŒ–ã‚’åˆ†ã‹ã‚‰ãªã„ã‚ˆã†ã«ã—ãªãŒã‚‰è‰²ã‚’æ¸›ã‚‰ã—ã¦ã„ããŸã‚ã®ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ å®Ÿè£…ã§ã™ã€‚ï¼ˆã“ã®è¨˜äº‹ã§ã¯æ¸›è‰²ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ ã«ã¤ã„ã¦ã®èª¬æ˜Žã¯çœç•¥ã—ã¾ã™ã€‚ï¼‰ ãƒ†ãƒ¼ãƒ–ãƒ«ã‚’ä½œæˆã—ãŸã‚‰ã€ç”»åƒã®ãã‚Œãžã‚Œã®ãƒ”ã‚¯ã‚»ãƒ«ã‚’ RGB å½¢å¼ã‹ã‚‰ãƒ†ãƒ¼ãƒ–ãƒ«ã®ä½•ç•ªç›®ã®è‰²ã‚’ä½¿ã†ã‹ã«ç½®ãæ›ãˆã¾ã™ã€‚ ä¸Šå›³ã®ã‚ˆã†ã«ã€1 ãƒ”ã‚¯ã‚»ãƒ«ã‚ãŸã‚Š 24bit å¿…è¦ã ã£ãŸç”»åƒãŒ 1 ãƒ”ã‚¯ã‚»ãƒ«ã‚ãŸã‚Š 8bit ã«ãªã£ãŸã®ã§ã€ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚ºã¯å¤§ä½“ 1/3 ã«ãªã‚Šã¾ã™ã€‚ ï¼ˆãƒ‘ãƒ¬ãƒƒãƒˆã®ãƒ‡ãƒ¼ã‚¿ã«æœ€å¤§ 3 Byte * 256 = 768 Byte å¿…è¦ã¨ã‹ã€åŒã˜ã‚ˆã†ã«åœ§ç¸®ã•ã‚Œãªã„ã¨
pcod 2012/11/06
algorithm
ãƒªãƒ³ã‚¯
Fast Data in the Era of Big Data: Twitter's Real-Time Related Query Suggestion Architecture
We present the architecture behind Twitter's real-time related query suggestion and spelling correction service. Although these tasks have received much attention in the web search literature, the Twitter context introduces a real-time "twist": after significant breaking news events, we aim to provide relevant results within minutes. This paper provides a case study illustrating the challenges of
pcod 2012/11/05
Twitter

search

algorithm

pdf
ãƒªãƒ³ã‚¯
GitHub - komiya-atsushi/lz4-ruby: Ruby bindings for LZ4 (Extremely Fast Compression algorithm)
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
pcod 2012/07/06
LZ4 for ruby ã™ã’ãˆã“ã‚“ãªgemã‚ã‚‹ã‚“ã

ruby

algorithm
ãƒªãƒ³ã‚¯
Sux: Implementing Succinct Data Structures
Sux in an umbrella nickname for the results of my fiddling with the implementation of basic succinct data strucures. The resulting code is rather sparse. The main highlights are: a novel, broadword-based implementation of rank/select queries for up to 264 bits that is highly competitive with known 32-bit implementations on 64-bit architectures (additional space required is 25% for ranking and 12.
pcod 2012/04/20
java

library

algorithm
ãƒªãƒ³ã‚¯
Home
Created for the August 2024 Artcrush Gallery curation by Kika Nicolela
pcod 2012/02/13
processing

algorithm

art
ãƒªãƒ³ã‚¯
Redis bitmaps - Fast, easy, realtime metrics -
At Spool, we calculate our key metrics in real time. Traditionally, metrics are performed by a batch job (running hourly, daily, etc.). Redis backed bit maps allow us to perform such calculations in realtime and are extremely space efficient. In a simulation of 128 million users, a typical metric such as â€œdaily unique usersâ€ takes less than 50 ms on a MacBook Pro and only takes 16 MB of memory. Spo
pcod 2011/12/02
algorithm

redis
ãƒªãƒ³ã‚¯
æœ€è¿‘å‚æŽ¢ç´¢2011 - Preferred Networks Tech Blog
ã“ã‚“ã«ã¡ã¯ã€äºŒå°ç›®ã®mbaã‚’è²·ã†ã®ã‚’ãŸã‚ã‚‰ã£ã¦ã„ã‚‹å²¡é‡ŽåŽŸã§ã™ã€‚ ã‚¢ã‚¤ãƒ†ãƒ é›†åˆã«å¯¾ã—ã€ä¸Žãˆã‚‰ã‚ŒãŸã‚¢ã‚¤ãƒ†ãƒ ã¨ä¼¼ã¦ã„ã‚‹ã‚¢ã‚¤ãƒ†ãƒ ã‚’æ±‚ã‚ã‚‹ã€ã¨ã„ã†è¿‘å‚æŽ¢ç´¢å•é¡Œã¯å¤å…¸çš„ãªå•é¡Œã§ã‚ã‚ŠãªãŒã‚‰ã€ç¾åœ¨ã§ã‚‚å¤šãã®æ”¹å–„ãŒã•ã‚Œã¦ã„ã¾ã™ã€‚ç‰¹ã«è¨€èªžæƒ…å ±ã€ç”»åƒæƒ…å ±ã€è¡Œå‹•å±¥æ´æƒ…å ±ã€ç”Ÿç‰©æƒ…å ±ã€è³¼è²·æƒ…å ±ãªã©ã‚ã‚Šã¨ã‚ã‚‰ã‚†ã‚‹ãƒ‡ãƒ¼ã‚¿ãŒé«˜æ¬¡å…ƒä¸ã®ç‚¹ã¨ã—ã¦è¡¨ç¾ã•ã‚Œã‚‹ã‚ˆã†ã«ãªã£ã¦ãã¦ãŠã‚Šã€ã“ã†ã—ãŸãƒ‡ãƒ¼ã‚¿ã®æœ€è¿‘å‚æŽ¢ç´¢ã¯åºƒã„åˆ†é‡Žã§å¿œç”¨ç¯„å›²ãŒã‚ã‚‹æŠ€è¡“ã«ãªã£ã¦ã„ã¾ã™ã€‚ ã‚¢ã‚¤ãƒ†ãƒ ãŒä½Žæ¬¡å…ƒï¼ˆä¾‹ãˆã°2, 3æ¬¡å…ƒï¼‰ã®å ´åˆã¯kdæœ¨ã‚„æœ€è¿‘ã ã¨waveletæœ¨ã‚’ä½¿ã†æ–¹æ³•ãŒã‚ã‚Šã¾ã™ãŒã€ä»Šå›žæ‰±ã†ã‚±ãƒ¼ã‚¹ã¯å„ã‚¢ã‚¤ãƒ†ãƒ ãŒé«˜æ¬¡å…ƒï¼ˆæ•°ç™¾ä¸‡æ¬¡å…ƒï¼‰ä¸ã®ç‚¹ã§ã‚ã£ãŸã‚Šã€ã‚¢ã‚¤ãƒ†ãƒ é–“ã®è·é›¢ã®ã¿ãŒå®šç¾©ã•ã‚Œã¦ã„ã‚‹å ´åˆï¼ˆã‚«ãƒ¼ãƒãƒ«é–¢æ•°ãªã©ï¼‰ã§ã™ã€‚ã‚¢ã‚¤ãƒ†ãƒ æ•°ã¯æ•°ä¸‡ã‹ã‚‰æ•°å„„ãã‚‰ã„ã‚’æƒ³å®šã—ã¦ã„ã¾ã™ã€‚ æœ€è¿‘å‚æŽ¢ç´¢å•é¡Œã¯ã„ãã¤ã‹ã‚ã‚Šã¾ã™ãŒã€ä¾‹ãˆã°kè¿‘å‚ã‚°ãƒ©ãƒ•æ§‹ç¯‰å•é¡Œã§ã¯ã€ ã€Œã‚¢ã‚¤ãƒ†ãƒ é›†åˆX = x1,
pcod 2011/08/05
algorithm
ãƒªãƒ³ã‚¯
dag_vector: ãƒ©ãƒ³ãƒ€ãƒ ã‚¢ã‚¯ã‚»ã‚¹å¯èƒ½ãªåœ§ç¸®é…åˆ— - Preferred Networks Tech Blog
ã“ã‚“ã«ã¡ã¯ã€ã“ã®å¤ã¯ã‚·ãƒ«ã‚ãƒ¼ãƒ‰ãƒ©ã‚¤ã§ä¹—ã‚Šåˆ‡ã‚ŠãŸã„å²¡é‡ŽåŽŸã§ã™ã€‚ ä»Šæ—¥ã¯æœ€è¿‘å…¬é–‹ã—ãŸC++ã®ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹ã§ã‚ã‚‹dag vectorã«ã¤ã„ã¦ç´¹ä»‹ã—ã¾ã™ã€‚ github: dag_vector ãƒ©ã‚¤ã‚»ãƒ³ã‚¹ã¯ä¿®æ£BSDãƒ©ã‚¤ã‚»ãƒ³ã‚¹ã§ã™ã€‚ dag vector (direct accessible gamma code vector) ã¯å€¤ã‚’åœ§ç¸®ã—ã¦æ ¼ç´ã—ãŸã¾ã¾ä»»æ„ã®å ´æ‰€ã®å€¤ã‚’é«˜é€Ÿã«å‚ç…§å¯èƒ½ãªé…åˆ—ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã§ã™ã€‚ã—ã‹ã‚‚ãƒ‡ãƒ¼ã‚¿æœ«å°¾ã¸ã®è¿½è¨˜ãŒå¯èƒ½ã§ã™ã€‚ dag vectorã¯std::vectorã®ã‚ˆã†ã«åˆ©ç”¨ã§ãã¾ã™ã€‚ä¸‹ã«ã„ãã¤ã‹ä¾‹ã‚’è¦‹ã¦ã„ãã¾ã—ã‚‡ã†ã€‚ dag_vectorã®ä¾‹ #include "dag_vector.hpp" // dag_vectorã¯0ä»¥ä¸Šã®æ£æ•´æ•°ã®é…åˆ—ã‚’æ‰±ã†é…åˆ—ã€‚ dag_vector dv; // å€¤ã¯ã„ã¤ã§ã‚‚è¿½åŠ å¯èƒ½ã€‚è¿½åŠ ã•ã‚ŒãŸå€¤ã¯åœ§ç¸®ã—ã¦æ ¼ç´ã•ã‚Œã‚‹ // æ£æ•´æ•°xã¯2lg(
pcod 2011/06/27
algorithm

library
ãƒªãƒ³ã‚¯
projects:sgd [leon.bottou.org]
Learning algorithms based on Stochastic Gradient approximations are known for their poor performance on optimization tasks and their extremely good performance on machine learning tasks (Bottou and Bousquet, 2008). Despite these proven capabilities, there were lingering concerns about the difficulty of setting the adaptation gains and achieving robust performance. Stochastic gradient algorithms ha
pcod 2011/06/26
algorithm

tool

nlp

svm

machine learning
ãƒªãƒ³ã‚¯
æµ·é¢¨ã«æºã‚Œã‚‹ä¸€è¼ªã®TRIEãƒ©ã‚¤ãƒ–ãƒ©ãƒª erikaã‚’ä½œã£ã¦ã¿ãŸã‚ˆ - EchizenBlog-Zwei
æ³¨æ„ï¼šã“ã®è¨˜äº‹ã®å†…å®¹ã¯å¤ã„ã‚‚ã®ã§ã™ã€‚ æœ€æ–°ç‰ˆã®erika-trieã¯ erika-trie(å®Ÿç”¨ç‰ˆ)ã¨ã‚ãƒ¼ãƒ¯ãƒ¼ãƒ‰æŠ½å‡ºãƒ„ãƒ¼ãƒ«erika_extractã‚’ä½œã£ãŸã‚ˆ - EchizenBlog-Zwei ã†ã£ã‹ã‚Šæ‰‹ãŒæ»‘ã£ã¦è‡ªåˆ†ã§â˜†ã¤ã‘ã¦ã—ã¾ã„ã¾ã—ãŸãŒè‡ªç”»è‡ªè³›ã—ã¦ã‚‹ã‚ã‘ã§ã¯ãªã„ã§ã™ãƒ»ãƒ»ãƒ»ã€‚ ã‚’ã”å‚ç…§ãã ã•ã„ã€‚ æœ€è¿‘TRIE(ãƒˆãƒ©ã‚¤)ã«å¯¾ã™ã‚‹èˆˆå‘³ãŒé«˜ã¾ã£ã¦ããŸã®ã§å‹‰å¼·ã®ãŸã‚ã«ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã‚’ä½œã£ã¦ã„ãŸã€‚ä¸€å¿œå®Œæˆã—ãŸã®ã§å…¬é–‹ã—ã¦ãŠãã€‚æ™®é€šã®TRIEã¨LOUDSã«ã‚ˆã‚‹TRIEã®äºŒã¤ã‚’å®Ÿè£…ã—ãŸã€‚ åå‰ã¯erikaã€‚ã‚‚ã—ãã¯erika-trieã€‚ä»¥å‰ä½œã£ãŸCompressed Suffix Arrayã®ãƒ©ã‚¤ãƒ–ãƒ©ãƒªãŒtsubomiã ã£ãŸã®ã§ã€ä»Šå›žã¯erikaã«ã—ãŸã€‚ã‚³ãƒ¼ãƒ‰ã¯GoogleCodeã®tsubomiã¨åŒã˜ã¨ã“ã‚ã«ãŠã„ã¦ã‚ã‚‹ã€‚ã¨ã„ã†ã‹tsubomiã®ä»˜å±žå“çŠ¶æ…‹ã€‚é–‹ç™ºã¯linuxã§è¡Œã£ãŸãŒç‰¹ã«ç’°å¢ƒã«ä¾å˜ã—ãŸ
pcod 2011/05/26
algorithm
ãƒªãƒ³ã‚¯
SDM2011ã§waveletæœ¨ã‚’ç”¨ã„ãŸå¤§è¦æ¨¡ã‚°ãƒ©ãƒ•ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã®é«˜é€Ÿé¡žä¼¼åº¦æ¤œç´¢æ‰‹æ³•ã«ã¤ã„ã¦ç™ºè¡¨ã—ã¾ã—ãŸ - Yasuo Tabeiã®æ—¥è¨˜
4æœˆ28æ—¥ã‹ã‚‰4æœˆ30æ—¥ã«é–‹å‚¬ã•ã‚ŒãŸãƒ‡ãƒ¼ã‚¿ãƒ¼ãƒžã‚¤ãƒ‹ãƒ³ã‚°ã®å›½éš›ä¼šè° SIAM Conference on Data Mining (SDM2011)ã«ã¦waveletæœ¨ã‚’ç”¨ã„ãŸå¤§è¦æ¨¡ã‚°ãƒ©ãƒ•ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã®é«˜é€Ÿé¡žä¼¼åº¦æ¤œç´¢æ‰‹æ³•ã«ã¤ã„ã¦ç™ºè¡¨ã—ã¦ãã¾ã—ãŸã€‚ Yasuo Tabei and Koji Tsuda: Kernel-based Similarity Search in Massive Graph Databases with Wavelet Trees, Eleventh SIAM International Conference on Data Mining (SDM), Arizona, USA, 2011. Link to the paper æœ¬ç ”ç©¶ã§ã¯å¤§è¦æ¨¡ã‚°ãƒ©ãƒ•ãƒ‡ãƒ¼ã‚¿ãƒ¼ã®ç´¢å¼•ã«ã‚ˆã‚‹é«˜é€Ÿãªé¡žä¼¼åº¦æ¤œç´¢æ‰‹æ³•ã‚’ææ¡ˆã—ã¾ã—ãŸã€‚è¿‘å¹´ã€ã‚°ãƒ©ãƒ•ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã«ç™»éŒ²ã•ã‚Œã¦ã„ã‚‹ã‚°ãƒ©ãƒ•ã®æ•°ã¯å¤§è¦æ¨¡åŒ–ã—ã¦ã„
pcod 2011/05/09
algorithm
ãƒªãƒ³ã‚¯
Mneme: Scalable Duplicate Filtering Service - igvita.com
By Ilya Grigorik on March 24, 2011 Detecting and dealing with duplicates is a common probl em: sometimes we want to avoid performing an operation based on this knowledge, and at other times, like in a case of a database, we want may want to only permit an operation based on a hit in the filter (ex: skip disk access on a cache miss). How do we build a system to solve the probl em? The solution will d
pcod 2011/03/27
redis

algorithm
ãƒªãƒ³ã‚¯
Showing Relevant Ads via Lipschitz Context Multi-Armed Bandits
pcod 2011/03/21
We want to design an online algorithm, which given a query in each time step and a history of past queries and ad clicks, displays an ad to maximize the expected number of clicks.

algorithm

memo
ãƒªãƒ³ã‚¯
1 2 3 4 5 6 7 8 9 10 æ¬¡ã®ãƒšãƒ¼ã‚¸