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

yuroyoro id:yuroyoro

æ©Ÿæ¢°å¦ç¿’ã«é–¢ã™ã‚‹yuroyoroã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (12)

${{author_name}}$

{{{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€ä¸–ç•Œä¸€åˆ†ã‹ã‚Šã‚„ã™ã„è§£èª¬ã€‘ã‚¤ãƒ©ã‚¹ãƒˆã§ã¿ã‚‹TransformerBeginaid
æœ¬è¨˜äº‹ã¯The Illustrated Transf ormerã‚’å’Œè¨³ã—ãŸå†…å®¹ã«ãªã‚Šã¾ã™ã€‚å¼•ç”¨å…ƒã¯Jay Alammarã•ã‚“ï¼ˆ@JayAlammarï¼‰ãŒåŸ·ç†ã•ã‚ŒãŸãƒ–ãƒã‚°è¨˜äº‹ã§ï¼ŒMITã®æŽˆæ¥ã§ã‚‚å®Ÿéš›ã«åˆ©ç”¨ã•ã‚Œã¦ã„ã¾ã™ã€‚ æ‰€ã€…ã«ç®¡ç†äººã®æ³¨é‡ˆãŒå…¥ã‚Šã¾ã™ã€‚ãã®å ´åˆã¯ï¼Œã“ã®ã‚ˆã†ãªé‰›ç†å°ã®ãƒœãƒƒã‚¯ã‚¹å†…ã«è¨˜è¿°ã™ã‚‹ã‚ˆã†ã«ã—ã¦ã„ã¾ã™ã€‚ã‚‚ã—ç¿»è¨³é–“é•ã„ãªã©ãŒã”ã–ã„ã¾ã—ãŸã‚‰ï¼Œã”æŒ‡æ‘˜ã„ã„ãŸã ã‘ã¾ã™ã¨å¹¸ã„ã§ã™ã€‚ãªãŠï¼Œæœ¬è¨˜äº‹ã®ä½œæˆã«é–¢ã—ã¦ã¯ï¼ŒJay Alammarã•ã‚“ã«è¨±å¯ã‚’ã„ãŸã ã„ã¦ãŠã‚Šã¾ã™ã€‚ ã¯ã˜ã‚ã« å‰å›žã®è¨˜äº‹ã§ã¯ï¼Œæ³¨æ„æ©Ÿæ§‹ã«ã¤ã„ã¦ãŠä¼ãˆã—ã¾ã—ãŸã€‚æ³¨æ„æ©Ÿæ§‹ã¨ã¯ï¼Œç¾ä»£ã®æ·±å±¤å¦ç¿’ã«ãŠã„ã¦è‡³ã‚‹æ‰€ã§åˆ©ç”¨ã•ã‚Œã¦ã„ã‚‹æ‰‹æ³•ã§ï¼Œãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«æ©Ÿæ¢°ç¿»è¨³ã®ç²¾åº¦å‘ä¸Šã«å¤§ããè²¢çŒ®ã—ãŸæ¦‚å¿µã§ã™ã€‚æœ¬è¨˜äº‹ã§ã¯ï¼Œæ³¨æ„æ©Ÿæ§‹ã‚’åˆ©ç”¨ã—ã¦ãƒ¢ãƒ‡ãƒ«ã®å¦ç¿’é€Ÿåº¦ã‚’å‘ä¸Šã•ã›ã‚‹Transf ormerã«ã¤ã„ã¦è¦‹ã¦ã„ãã¾ã—ã‚‡ã†ã€‚Transf ormerã¯ç‰¹å®šã®ã‚¿ã‚¹ã‚¯ã«ãŠã„ã¦ï¼ŒG
yuroyoro 2023/04/04
æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
ã€ã‚¹ãƒžãƒ›OK/å®Ÿè¡Œã—ãªãŒã‚‰å¦ã¶ã€‘æ±å¤§æ¾å°¾ç ”ã®ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆè‚²æˆ/ Deep LearningåŸºç¤Žè¬›åº§ã‚’è‡ªç¿’ã™ã‚‹ - Qiita
2. numpy, scipy, scikit-learn ã®ä½¿ã„æ–¹ã‚’ç†è§£ã™ã‚‹ 3. k-Nearest Neighbors (k-NN) ã‚’ä½¿ã£ãŸæ‰‹æ›¸ãæ–‡å—èªè˜ 4. ãƒã‚¸ã‚¹ãƒ†ã‚£ãƒƒã‚¯å›žå¸°ã®å®Ÿè£…ã¨å¦ç¿’ã€æ´»æ€§åŒ–é–¢æ•°ã¨ãã®å¾®åˆ†ã®å®Ÿè£…ã€å¤šå±¤ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³ã®å®Ÿè£…ã¨å¦ç¿’ 5. Tensorflowã®åŸºç¤Žã‚’å¦ã¶ 6. Denoising Autoencoderã®å®Ÿè£…. ã¾ãŸ, MNISTã‚’ç”¨ã„ã¦æ¬¡ã®ã“ã¨ã‚’ç¢ºèªã€Stacked Denoising Autoencoder (SdA) ã®å®Ÿè£… 7. ç•³ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯(Convolutional Neural Networks)ã®å®Ÿè£…ã¨å¦ç¿’ 8. CIFAR10ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã‚’ä½¿ã£ãŸAugmentationã€å‰å‡¦ç†ã€Batch Normalizationã€CNNå®Ÿè£…ã€Activationå¯è¦–åŒ– 9. Recurrent Neural Networ
yuroyoro 2020/05/29
æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
Goè¨€èªžã§æ‰±ãˆã‚‹ãƒ‡ãƒ¼ã‚¿ãƒ•ãƒ¬ãƒ¼ãƒ åŽ³é¸4ã¤ - Qiita
ã¯ã˜ã‚ã« ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã§ãªã‹ã£ãŸã¨ã—ã¦ã‚‚ã€æ•°å€¤ãƒ‡ãƒ¼ã‚¿ã‚’ä½¿ã£ã¦æ§˜ã€…ãªè§£æžã‚’ã™ã‚‹éš›ã«ã¯ CSV ãƒ•ã‚¡ã‚¤ãƒ«ç‰ãƒ•ã‚¡ã‚¤ãƒ«ã‚’èªã¿è¾¼ã¿ã€æ•°å€¤ã®é…åˆ—ã¨ã—ã¦ãƒ¡ãƒ¢ãƒªã«ä¿æŒã—ã¦ã€ãã‚Œã‚‰ã‚’ãƒ«ãƒ¼ãƒ—ç‰ã§åˆ©ç”¨ã—ã¦è§£æžã‚’è¡Œã£ã¦ãŠã‚‰ã‚Œã‚‹ã¨æ€ã„ã¾ã™ã€‚ ãã®éš›ã€é…åˆ—ã¯1æ¬¡å…ƒç›®ã«è¡Œã€2æ¬¡å…ƒç›®ã«åˆ—ã€ã‚’æ ¼ç´ã™ã‚‹ã®ãŒä¸€èˆ¬çš„ã§ã™ã€‚å¤šãã®ã‚±ãƒ¼ã‚¹ã§ã¯ã“ã®æ–¹æ³•ã§äº‹è¶³ã‚Šã‚‹ã®ã§ã™ãŒã€è§£æžã‚’è¡Œã†ã†ã¡ã«ã€Œåˆ—ã¨ã—ã¦ãƒ‡ãƒ¼ã‚¿ã®å›ºã¾ã‚Šã‚’æ‰±ã„ãŸã„ã€ã€Œãƒ©ãƒ™ãƒ«ä»˜ã‘ã•ã‚ŒãŸåˆ—ã‚’æ‰±ã„ãŸã„ã€ã¨æ„Ÿã˜ã‚‹äº‹ãŒå‡ºã¦ãã‚‹ã¨æ€ã„ã¾ã™ã€‚ ã“ã‚Œã‚’ç°¡å˜ã«ã—ã¦ãã‚Œã‚‹ã®ãŒã€Œãƒ‡ãƒ¼ã‚¿ãƒ•ãƒ¬ãƒ¼ãƒ ã€ã§ã™ã€‚ ãƒ‡ãƒ¼ã‚¿ãƒ•ãƒ¬ãƒ¼ãƒ 4ç¨® æœ¬è¨˜äº‹ã§ã¯ Go è¨€èªžã‹ã‚‰æ‰±ãˆã‚‹ãƒ‡ãƒ¼ã‚¿ãƒ•ãƒ¬ãƒ¼ãƒ ã‚’4ã¤ã”ç´¹ä»‹ã—ã¾ã™ã€‚ QFrame https://github.com/tobgu/qframe QFrame ã¯ã€ãƒ•ã‚£ãƒ«ã‚¿ãƒªãƒ³ã‚°ã€é›†è¨ˆã€ãŠã‚ˆã³ãƒ‡ãƒ¼ã‚¿æ“ä½œã‚’ã‚µãƒãƒ¼ãƒˆã™ã‚‹ã‚¤ãƒŸãƒ¥ãƒ¼ã‚¿ãƒ–ãƒ«ãªãƒ‡ãƒ¼ã‚¿ãƒ•ãƒ¬ãƒ¼ãƒ ã§ã™ã€‚ QFrame ã§ã®
yuroyoro 2019/12/18
GoLang

æ©Ÿæ¢°å¦ç¿’

ML

DataFrame
ãƒªãƒ³ã‚¯
GitHub - matsuolab-edu/dl4us
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
yuroyoro 2019/05/16
DeepLearning

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
ãƒ©ãƒ³ãƒ€ãƒ ãƒ•ã‚©ãƒ¬ã‚¹ãƒˆã¨æ¤œå®šã‚’ç”¨ã„ãŸç‰¹å¾´é‡é¸æŠžæ‰‹æ³• Boruta - å¦ç¿’ã™ã‚‹å¤©ç„¶ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆ
ç‰¹å¾´é‡é¸æŠžã¨ã¯ Borutaã¨ã¯ ã¨ã‚Šã‚ãˆãšä½¿ã£ã¦ã¿ã‚‹ ãƒ™ãƒ¼ã‚¹ãƒ©ã‚¤ãƒ³ã®åˆ¤åˆ¥ Borutaã®åˆ¤åˆ¥ Borutaã®ã‚¢ã‚¤ãƒ‡ã‚¢ã®æ¦‚è¦ Borutaã®ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ 1. åˆ¤åˆ¥ã«å¯„ä¸Žã—ãªã„ã¯ãšã®å½ã®ç‰¹å¾´é‡ã‚’ä½œã‚‹ã€‚ 2. å½ã®ç‰¹å¾´é‡ã¨ä¸€ç·’ã«ãƒ©ãƒ³ãƒ€ãƒ ãƒ•ã‚©ãƒ¬ã‚¹ãƒˆã‚’è¨“ç·´ã€‚ 3. å„ç‰¹å¾´é‡ã®é‡è¦åº¦ã¨å½ã®ç‰¹å¾´é‡ã®ç‰¹å¾´é‡ã‚’æ¯”è¼ƒã€‚ 4. è¤‡æ•°å›žæ¯”è¼ƒã—æ¤œå®šã‚’è¡Œã†ã“ã¨ã§ã€æœ¬å½“ã«é‡è¦ãªç‰¹å¾´é‡ã®ã¿ã‚’é¸æŠžã€‚ æ¤œå®šã«ã¤ã„ã¦ 1. æ£„å´ã—ãŸã„å¸°ç„¡ä»®èª¬ã¨å—å®¹ã—ãŸã„å¯¾ç«‹ä»®èª¬ã‚’ç”¨æ„ã™ã‚‹ã€‚ 2. è¦³æ¸¬å€¤ã‹ã‚‰æ¤œå®šçµ±è¨ˆé‡Tã‚’å®šã‚ã‚‹ã€‚ 3. å¸°ç„¡ä»®èª¬ãŒæ£ã—ã„ã¨ã—ã¦Tã®åˆ†å¸ƒã‚’æ±‚ã‚ã‚‹ã€‚ 4. ååˆ†å°ã•ã„æœ‰æ„æ°´æº–Î±ã‚’å®šã‚ã€å¸°ç„¡ä»®èª¬ãŒæ£ã—ã„ã¨ãã«ã¨ãªã‚‹é ˜åŸŸã‚’æ£„å´åŸŸã¨ã™ã‚‹ã€‚ 5. è¦³æ¸¬ã•ã‚ŒãŸTãŒã«å…¥ã£ã¦ã„ãŸã‚‰å¯¾ç«‹ä»®èª¬ã‚’å—å®¹ã—ã€å…¥ã£ã¦ã„ãªã‘ã‚Œã°å¸°ç„¡ä»®èª¬ã‚’å—å®¹ã™ã‚‹ã€‚ ã¾ã¨ã‚ è£œè¶³ ä½¿ã†éš›ã®Tipsç‰ 2019/01/06è¿½è¨˜ å‚è€ƒ ç‰¹å¾´é‡é¸æŠžã¨ã¯ ç‰¹å¾´é‡é¸æŠž
yuroyoro 2019/01/05
æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
ã‚ªãƒ³ãƒ©ã‚¤ãƒ³åºƒå‘Šã«ãŠã‘ã‚‹ä¸æ£ã‚¯ãƒªãƒƒã‚¯æ¤œå‡ºæ‰‹æ³•ã¨æ´å²
2016-09-03 ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°+WEBæ±äº¬ã§ã®ç™ºè¡¨è³‡æ–™ã§ã™
yuroyoro 2016/09/05
AdTech

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
TensorFlow Playground ã§ã‚ã‹ã‚‹ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯
ã“ã®è¨˜äº‹ã‚’èªã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã«èˆˆå‘³ã‚’æŒã¡å‹‰å¼·ã‚’å§‹ã‚ã¾ã—ãŸã€‚ ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚’ç›®ã§è¦‹ã¦ç›´æ„Ÿçš„ã«ç†è§£ã§ãã‚‹ã®ã¯ç´ æ™´ã‚‰ã—ã„ã§ã™ãã€‚ åŠå¹´ä»¥ä¸Šã‚‚å‰ã®è¨˜äº‹ãªã®ã§ã‚³ãƒ¡ãƒ³ãƒˆã‚’èªã¾ã‚Œã¦ã„ã‚‹ã‹åˆ†ã‹ã‚Šã¾ã›ã‚“ãŒã€ãŠèžãã—ãŸã„ã“ã¨ãŒã‚ã‚Šã¾ã™ã€‚ TensorFlow Playgroundã®å…¥åŠ›å±¤ã®Featureã«ã¤ã„ã¦å‹‰å¼·ã‚’ã—ã¦ã„ã‚‹ã®ã§ã™ãŒã€ãªãœç”Ÿã®å…¥åŠ›å€¤(åº§æ¨™x,y)ã§ã¯ãªãã€Featureã‚’é–“ã«ã‹ã¾ã›ã¦ã„ã‚‹ã‹ãã®èƒŒæ™¯ã‚’æ•™ãˆã¦ã„ãŸã ããŸã„ã§ã™ã€‚ ã¨ã„ã†ã®ã‚‚ã€MNISTã®ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ç‰ã§ã¯ç¸¦æ¨ª28x28ãƒ”ã‚¯ã‚»ãƒ«ã®784å€‹ã®ã‚¢ãƒ‰ãƒ¬ã‚¹ã®ã‚°ãƒ¬ãƒ¼ã‚¹ã‚±ãƒ¼ãƒ«ã®å€¤ã‚’å…¥åŠ›ã¨ã—ã€å…¥åŠ›å±¤ã«784å€‹ã®ãƒ‹ãƒ¥ãƒ¼ãƒãƒ³ã‚’ä¸¦ã¹ã¦ã„ã‚‹è§£èª¬ãŒ å¤šãã€TensorFlow Playgroundã§è¡Œã‚ã‚Œã¦ã„ã‚‹ã‚ˆã†ãªFeatureã®é¸æŠžãŒã©ã“ã‹ã‚‰å‡ºã¦ããŸã‚‚ã®ãªã®ã‹åˆ†ã‹ã‚‰ãªã„ã®ã§ã™ã€‚ ã¡ã‚‡ã£ã¨å®£ä¼ã®ã‚ˆã†ã«ãªã£ã¦ã—ã¾ã„ã¾ã™ãŒã€è‡ªåˆ†ã®å¦ç¿’æˆæžœã®
yuroyoro 2016/07/20
DeepLearning

TensorFlow

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
Clojure/Incanter/clj-mlã§ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°å…¥é–€ - ã‚ã‚“ã¡ã¹ï¼
æ¦‚è¦ Clojureã§ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°ã«å¿…è¦ãªå„æ‰‹æ³•ã‚’è§£èª¬ã™ã‚‹è¨˜äº‹ã§ã™ã€‚ æœ¬è¨˜äº‹ã‚’èªã‚€ã¨ã€å…¨ãClojureã‚’çŸ¥ã‚‰ãªã„æ–¹ã§ã‚‚ ãƒ‡ãƒ¼ã‚¿æŠ½å‡ºãƒ»é›†è¨ˆ å¯è¦–åŒ– æ©Ÿæ¢°å¦ç¿’ï¼ˆæ±ºå®šæœ¨ã€ãƒ©ãƒ³ãƒ€ãƒ ãƒ•ã‚©ãƒ¬ã‚¹ãƒˆã€k-meansã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚°) ã‚’Clojureã§å®Ÿè¡Œã§ãã‚‹ã‚ˆã†ã«ãªã‚Šã¾ã™ã€‚ ã¯ã˜ã‚ã« Clojureã¨ã¯JVMä¸Šã§å‹•ãï¼ˆã¤ã¾ã‚ŠOSã‚’å•ã‚ãšæ²¢å±±ã®ç’°å¢ƒã§å‹•ãä¸Šã« Javaè³‡ç”£ã‚’ãã®ã¾ã¾ä½¿ãˆã‚‹ï¼‰Lispç³»ã®è¨€èªžã§ã™ã€‚ Clojureã§ã¯ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°ã‚’è¡Œã†éš›ã€ Incanterã¨clj-mlã¨ã„ã†2ã¤ã®ä¾¿åˆ©ãªãƒ„ãƒ¼ãƒ«ãŒã‚ã‚Šã¾ã™ã€‚ Incanterã¯çµ±è¨ˆå‡¦ç†ç”¨ã®å°‚é–€ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã§ã€ å˜ä½“ã§æ§˜ã€…ãªçµ±è¨ˆåˆ†æžã‚’è¡Œã£ãŸã‚Šé›†è¨ˆã‚’ä¾¿åˆ©ã«ã—ãŸã‚Šã™ã‚‹æ©Ÿèƒ½ãŒæ²¢å±±ç”¨æ„ã•ã‚Œã¦ã„ã¾ã™ã€‚ clj-mlã¯wekaã¨ã„ã†æ©Ÿæ¢°å¦ç¿’ç³»ã®ãƒ„ãƒ¼ãƒ«ã‚’Clojureã§ç°¡å˜ã«ä½¿ãˆã‚‹ã‚ˆã†ã«ã—ãŸãƒ©ãƒƒãƒ‘ãƒ¼ã§ã€ æ±ºå®šæœ¨ã‚„ãƒ©ãƒ³ãƒ€ãƒ ãƒ•ã‚©ãƒ¬ã‚¹ãƒˆãªã©ã®åˆ†é¡žå™¨ã‚„ å„ç¨®ã‚¯
yuroyoro 2014/02/16
clojure

æ©Ÿæ¢°å¦ç¿’

ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°
ãƒªãƒ³ã‚¯
Webãƒ‡ãƒ¼ã‚¿åˆ†æžï¼†ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ã‚¹ã§å½¹ç«‹ã¤çµ±è¨ˆå¦ãƒ»æ©Ÿæ¢°å¦ç¿’ç³»ã®åˆ†æžæ‰‹æ³•10é¸ - éŠ€åº§ã§åƒããƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã®ãƒ–ãƒã‚°
è¿½è¨˜ 2016å¹´3æœˆã«ä»¥ä¸‹ã®è¨˜äº‹ã«ã‚ˆã£ã¦ã“ã®å†…å®¹ã¯updateã•ã‚Œã¦ã„ã¾ã™ã€‚ä»Šå¾Œã¯ãã¡ã‚‰ã‚’ãŠèªã¿ä¸‹ã•ã„ã€‚ ä¸»ã«è‡ªåˆ†å‘ã‘ã®ã¾ã¨ã‚ã¨ã„ã†æ„å‘³åˆã„ãŒå¼·ã„ã‚“ã§ã™ãŒï¼ˆç¬‘ï¼‰ã€åƒ•ãŒå®Ÿéš›ã«2013å¹´6æœˆç¾åœ¨webãƒ‡ãƒ¼ã‚¿åˆ†æžï¼†ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ã‚¹ã®å®Ÿå‹™ã§ãƒ„ãƒ¼ãƒ«ãƒ»ãƒ©ã‚¤ãƒ–ãƒ©ãƒªãƒ»ãƒ‘ãƒƒã‚±ãƒ¼ã‚¸ã‚’åˆ©ç”¨ã—ã¦ã„ã‚‹ã‚‚ã®ã«é™ã£ã¦ã€çµ±è¨ˆå¦ãƒ»æ©Ÿæ¢°å¦ç¿’ç³»ã®åˆ†æžæ‰‹æ³•ã‚’10å€‹æŒ™ã’ã¦ç´¹ä»‹ã—ã¦ã¿ã‚ˆã†ã¨æ€ã„ã¾ã™ã€‚ è¿½è¨˜ å›žå¸°åˆ†æžï¼ˆç‰¹ã«ç·šå½¢é‡å›žå¸°åˆ†æžï¼‰ ç‹¬ç«‹æ€§ã®æ¤œå®šï¼ˆã‚«ã‚¤äºŒä¹—æ¤œå®šãƒ»ãƒ•ã‚£ãƒƒã‚·ãƒ£ãƒ¼ã®æ£ç¢ºç¢ºçŽ‡æ¤œå®šï¼‰ ä¸»æˆåˆ†åˆ†æž(PCA) / å› ååˆ†æž ã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚° æ±ºå®šæœ¨ / å›žå¸°æœ¨ ã‚µãƒãƒ¼ãƒˆãƒ™ã‚¯ã‚¿ãƒ¼ãƒžã‚·ãƒ³(SVM) ãƒã‚¸ã‚¹ãƒ†ã‚£ãƒƒã‚¯å›žå¸° ãƒ©ãƒ³ãƒ€ãƒ ãƒ•ã‚©ãƒ¬ã‚¹ãƒˆ ã‚¢ã‚½ã‚·ã‚¨ãƒ¼ã‚·ãƒ§ãƒ³åˆ†æžï¼ˆãƒã‚¹ã‚±ãƒƒãƒˆåˆ†æžãƒ»ç›¸é–¢ãƒ«ãƒ¼ãƒ«æŠ½å‡ºï¼‰ è¨ˆé‡æ™‚ç³»åˆ—åˆ†æž ãŠã‚ã‚Šã« ãŠã¾ã‘1ï¼šã€Œç´ æ€§ãƒ™ã‚¯ãƒˆãƒ«ï¼‹åˆ†é¡žãƒ©ãƒ™ãƒ«ã€ãªã‚‹ãƒ‡ãƒ¼ã‚¿å‰å‡¦ç† ãŠã¾ã‘2ï¼šã‚°ãƒ©ãƒ•ç†è«–*10 {igraph}ãƒ‘ãƒƒã‚±ãƒ¼ã‚¸ã§ã‚°ãƒ©
yuroyoro 2013/06/11
æ©Ÿæ¢°å¦ç¿’

çµ±è¨ˆ

ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°
ãƒªãƒ³ã‚¯
æ©Ÿæ¢°å¦ç¿’ã®ç†è«–ã¨å®Ÿè·µ
SACSIS2013ã§ã®ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«è¬›æ¼”è³‡æ–™ã§ã™ã€‚æ©Ÿæ¢°å¦ç¿’ã®å°Žå…¥ï¼šèƒŒæ™¯ã€æ‰‹æ³•ã€ç†è«–ã€å¿œç”¨ï¼‰ã€å®Ÿè·µï¼šã‚ªãƒ³ãƒ©ã‚¤ãƒ³å¦ç¿’+ç·šå½¢åˆ†é¡žã§å®Ÿéš›ä½œã£ã¦ã¿ã‚‹ã€ä½¿ã†éš›ã®èª²é¡Œã€ç™ºå±•ï¼šåˆ†æ•£+ãƒªã‚¢ãƒ«ã‚¿ã‚¤ãƒ ã§ã®æ©Ÿæ¢°å¦ç¿’ï¼ˆJubatusï¼‰ã€æ·±å±¤å¦ç¿’ï¼ˆDeep Neural Netï¼‰ã«ã¤ã„ã¦ã¾ã¨ã‚ã¾ã—ãŸRead less
yuroyoro 2013/05/23
æ©Ÿæ¢°å¦ç¿’
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
yuroyoro 2013/01/10
æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
LIBLINEARã‚’ç”¨ã„ãŸæ©Ÿæ¢°å¦ç¿’å…¥é–€(å˜èªžåˆ†å‰²)
ã“ã®ãƒšãƒ¼ã‚¸ã§ã¯æ©Ÿæ¢°å¦ç¿’ã®ãƒ„ãƒ¼ãƒ«(LIBLINEAR)ã‚’åˆ©ç”¨ã—ã¦ã€å®Ÿéš›ã«åˆ†é¡žå•é¡Œã‚’è§£ãã«ã¯ã©ã†ã„ã†æ‰‹é †ã‚’çµŒã‚‹ã‹ã¨ã„ã†ã“ã¨ã«ã¤ã„ã¦è§£èª¬ã—ã¾ã™ã€‚ã¤ã¾ã‚Šã€Kytea(äº¬éƒ½ãƒ†ã‚ã‚¹ãƒˆè§£æžãƒ„ãƒ¼ãƒ«ã‚ãƒƒãƒˆ)ã«ãŠã‘ã‚‹ç°¡æ˜“ç‰ˆã®å˜èªžåˆ†å‰²ãƒ¢ãƒ‡ãƒ«ã‚’ä½œã£ã¦ã¿ã‚ˆã†ã¨ã„ã†ã“ã¨ã§ã™ã€‚ ãªãŠä»Šå›žã¯ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°è¨€èªžã¨ã—ã¦Rubyã‚’ç”¨ã„ã¾ã™ãŒã€Rubyã®çŸ¥è˜ãŒãªãã¦ã‚‚å®Ÿè£…ãŒã§ãã‚‹ã‚ˆã†ã«è§£èª¬ã™ã‚‹ã‚ˆã†å¿ƒãŒã‘ã¾ã™ã€‚ã¾ãŸã€å¿…è¦ä»¥ä¸Šã«ç´°ã‹ãæ›¸ã„ã¦ã‚ã‚‹ã‹ã‚‚ã—ã‚Œã¾ã›ã‚“ãŒã€ä¸è¦ãªæ–¹ã¯é©å®œèªã¿é£›ã°ã—ã¦ä¸‹ã•ã„ã€‚ ç´°ã‹ã„èª¬æ˜Žã¯ã™ã£ã¨ã°ã—ã¦ã‚„ã‚Šæ–¹ã‚’è¦‹ã‚‹ æ©Ÿæ¢°å¦ç¿’ã£ã¦ï¼Ÿ æœ±é·ºã®æœWiki ã€Œæ©Ÿæ¢°å¦ç¿’ã€ãŒã‚ã‹ã‚Šã‚„ã™ã„ã‹ã¨æ€ã„ã¾ã™ã€‚ ã²ã¨ã“ã¨ã§è¨€ã†ã¨ã€ã€Œè¨“ç·´ãƒ‡ãƒ¼ã‚¿ã‚’ä¸Žãˆã¦ãã“ã‹ã‚‰æ©Ÿæ¢°ã«å•é¡Œã®è§£ãæ–¹ã‚’å¦ã‚“ã§ã‚‚ã‚‰ã„ã€åˆ¥ã®å•é¡Œã‚’è§£ã„ã¦ã‚‚ã‚‰ã†ã“ã¨ã€ã§ã™ã€‚ æ•™å¸«ã‚ã‚Šå¦ç¿’ãƒ»æ•™å¸«ãªã—å¦ç¿’ æ©Ÿæ¢°å¦ç¿’ã¯å¤§ããã€Œæ•™å¸«ã‚ã‚Šã€ã¨ã€Œæ•™å¸«ãªã—ã€ã«åˆ†ã‹ã‚Œã¾ã™ã€‚ ã€Œæ•™å¸«ã‚ã‚Šå¦ç¿’ã€ã¨ã¯
yuroyoro 2011/05/30
ruby

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
1