[B! æ•™å¸«ãªã—å¦ç¿’] p_tanã®ãƒ

p_tan id:p_tan

æ•™å¸«ãªã—å¦ç¿’ã«é–¢ã™ã‚‹p_tanã®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}}

ãƒ‰ãƒ¡ã‚¤ãƒ³çŸ¥è˜ãªã—æ•™å¸«ãªã—å¦ç¿’ã‚’å®Ÿç¾ã—ãŸImage GPTã€ç”»åƒç”Ÿæˆã‚‚ã™ã”ã„! (ç”»åƒã®è¡¨ç¾å¦ç¿’2020å¤ç‰¹é›†1)
3ã¤ã®è¦ç‚¹ âœ”ï¸Â æœªçŸ¥ãƒ‰ãƒ¡ã‚¤ãƒ³çŸ¥è˜ãªã—æ•™å¸«ãªã—è¡¨ç¾å¦ç¿’æˆåŠŸã€ä¾¡å€¤ã‚ã‚‹å®Ÿè¨¼å®Ÿé¨“(PoC)ã€è¨ˆç®—é‡ã¯åº¦å¤–è¦– âœ”ï¸ ç”Ÿæˆãƒ¢ãƒ‡ãƒ«ã¨ã—ã¦ã‚‚é©šãã®ç”»åƒç”Ÿæˆèƒ½åŠ› âœ”ï¸ ç²å¾—ã—ãŸè¡¨ç¾ã‚’ç”¨ã„ãŸç”»åƒåˆ†é¡žã§SOTAæ€§èƒ½ Generative Pretraining from Pixels written byÂ Mark Chen,Â Alec Radford,Â Ilya SutskeverÂ (OpenAI) (Submitted on 17 Jun 2020) Comments: Accepted at ICML2020 Subjects:Â Computer Vision and Pattern Recognition (cs.CV) PaperÂ Â Official CodeÂ COMM Code ä»Šå›žã‹ã‚‰3å›žã®äºˆå®šã§ã€ãƒ©ã‚¤ã‚¿ãƒ¼æŒã¡è¾¼ã¿ç‰¹é›†ä¼ç”»ã€Œç”»åƒã®è¡¨ç¾å¦ç¿’2020å¤ã€ã¨é¡Œã—ã¦ã€æ•™å¸«ãªã—å¦ç¿’ã«ã‚ˆã‚‹å„ç¨®æ‰‹æ³•ã‚’ã”ç´¹ä»‹
p_tan 2020/07/10
å¦ç¿’

ç”»åƒ

æ•™å¸«ãªã—å¦ç¿’
ãƒªãƒ³ã‚¯
[æœ€æ–°è«–æ–‡/GAN] ãƒ©ãƒ™ãƒ«ãŒãªã„ãªã‚‰æŽ¨æ¸¬ã™ã‚Œã°ã„ã„ã€‚ - Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? è«–æ–‡ç´¹ä»‹ãƒ»ç”»åƒå¼•ç”¨ Google Brainã‚ˆã‚Š2019.3.6æå‡º https://arxiv.org/pdf/1903.02271v1.pdf æœ¬ç ”ç©¶ã®GANã®ç‰¹å¾´ã¨æˆæžœ FIDã‚¹ã‚³ã‚¢ï¼ˆä½Žã„æ–¹ãŒè‰¯ã„ï¼‰ ã‚¹ã‚³ã‚¢ï¼˜ï½žï¼™ã®é–“ã«ã‚ã‚‹ç¸¦ç·šã¯ãƒ™ãƒ¼ã‚¹ãƒ©ã‚¤ãƒ³ï¼ˆã™ã¹ã¦ãƒ©ãƒ™ãƒ«ä»˜ã‘ã•ã‚ŒãŸç”»åƒã‚’ä½¿ã£ãŸBigGANï¼‰ æœ¬ç ”ç©¶ã®æ–¹æ³•ï¼ˆ$S^3GAN$ï¼‰ã§ã¯ ãƒ©ãƒ™ãƒ«ä»˜ã‘ã•ã‚ŒãŸç”»åƒã¯å…¨ä½“ã®ãŸã£ãŸï¼‘ï¼%ã«ã‚‚é–¢ã‚ã‚‰ãšSOTAã§ã‚ã‚‹BigGANã¨åŒç‰ã®æ€§èƒ½ã«ãªã£ãŸ ã¾ãŸå…¨ä½“ã®ï¼’ï¼%ã‚’ãƒ©ãƒ™ãƒ«ä»˜ã‘ã•ã‚ŒãŸç”»åƒã«ã™ã‚‹ã¨BigGANã‚’è¶…ãˆã‚‹æ€§èƒ½ã¨ãªã£ãŸ è§£åƒåº¦128Ã—128 ä¸Šæ®µ
p_tan 2019/03/12
GAN

è«–æ–‡

è§£èª¬

æ•™å¸«ãªã—å¦ç¿’

åŠæ•™å¸«ã‚ã‚Šå¦ç¿’
ãƒªãƒ³ã‚¯
UMAP è«–æ–‡ã‚’èªã‚€ - Qiita
ã“ã®è¨˜äº‹ã¯ æ©Ÿæ¢°å¦ç¿’ã®æ•°ç† Advent Calendar 2018 ã® 11 æ—¥ç›®ã®ãŸã‚ã®è¨˜äº‹ã§ã™. é–“ã«åˆã„ã¾ã—ãŸã‹? ã“ã®è¨˜äº‹ã«ã¤ã„ã¦ McInnes ã‚‰ã«ã‚ˆã‚‹ "Uniform Manifold Approximation and Projection (UMAP)" ã‚’èªã¿, ãã®è§£èª¬ã‚’è©¦ã¿ã¾ã™. UMAP ã¯ t-SNE ã®ã‚ˆã†ã«ãƒ‡ãƒ¼ã‚¿ã®æ¬¡å…ƒå‰Šæ¸›ã¨ãã®å¯è¦–åŒ–ã‚’æä¾›ã™ã‚‹æ‰‹æ³•ã§ã™. t-SNE ã‚ˆã‚Šã‚‚å¯è¦–åŒ–ã®çµæžœãŒã‚ˆã‚Šè‰¯ãè¦‹ãˆã‚‹ (åŒã˜ã‚¯ãƒ©ã‚¹ã‚¿ã¯ã‚ˆã‚Šã¾ã¨ã¾ã£ã¦ã„ã‚‹) ã“ã¨ã‚„, ãƒ‡ãƒ¼ã‚¿ã‚’ fuzzy topology ã¨ã„ã†ã“ã‚Œã¾ã§ä½¿ã‚ã‚Œã¦ã“ãªã‹ã£ãŸã‘ã©æ•°å¦çš„è£ä»˜ã‘ãŒã‚ã£ã¦ã‚¤ã‚±ã¦ã‚‹æ–¹æ³•ã§è¡¨ç¾ã—ãŸãœï¼ ã¨è¨€ã£ã¦ã‚‹ç‚¹ãŒç›®ã‚’å¼•ã, èªã‚“ã§ã¿ã‚ˆã†ã¨æ€ã£ãŸã®ã§ã™ãŒ, åœè«–ã®è¨€è‘‰ã§æ›¸ã‹ã‚Œã¦ãŠã‚Šèªã‚€ã®ã«è‹¦åŠ´ã—ã¦ã„ãŸã‚‰12æœˆã«ãªã£ã¦ã„ã¾ã—ãŸ. æ©Ÿæ¢°å¦ç¿’ç•Œéšˆã§åœè«–ã«é¦´æŸ“ã¿ãŒã‚ã‚‹ã“ã¨ã‚’ä»®å®šã™ã‚‹ã®ã¯ç„¡ç†ãŒ
p_tan 2018/12/12
æ©Ÿæ¢°å¦ç¿’

è«–æ–‡

ã‚ã¨ã§èªã‚€

UMAP

ã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚°

æ•™å¸«ãªã—å¦ç¿’

å¯è¦–åŒ–

å¤šæ§˜ä½“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“
p_tan 2018/05/01
ç¿»è¨³

æ©Ÿæ¢°å¦ç¿’

deep learning

deeplearning

è‡ªç„¶è¨€èªžå‡¦ç†

æ•™å¸«ãªã—å¦ç¿’
ãƒªãƒ³ã‚¯
20180115_æ±å¤§åŒ»å¦éƒ¨æ©Ÿèƒ½ç”Ÿç‰©å¦ã‚»ãƒŸãƒŠãƒ¼_æ·±å±¤å¦ç¿’ã®æœ€å‰ç·šã¨ã“ã‚Œã‹ã‚‰_å²¡é‡ŽåŽŸå¤§è¼”
2018å¹´1æœˆ15æ—¥ã®æ±äº¬å¤§å¦åŒ»å¦éƒ¨æ©Ÿèƒ½ç”Ÿç‰©å¦ã‚»ãƒŸãƒŠãƒ¼ã§ã®å²¡é‡ŽåŽŸå¤§è¼”ã®è¬›æ¼”è³‡æ–™ã§ã™ã€‚ æœ€å¾Œã®å‚è€ƒæ–‡çŒ®ã‚’ä¿®æ£ã—ã¾ã—ãŸã€‚ ä¿®æ£ç‰ˆã¯ã“ã¡ã‚‰ã§ã™ã€‚ https://www.slideshare.net/pfi/20180115-87025513 Read less
p_tan 2018/01/26
ãƒ‡ã‚£ãƒ¼ãƒ—ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°

deep learning

èª¤å·®é€†ä¼æ’

æ·±å±¤å¦ç¿’

deeplearning

æ•™å¸«ãªã—å¦ç¿’

ã‚¹ãƒ©ã‚¤ãƒ‰
ãƒªãƒ³ã‚¯
Deep Learning ã®æ¬¡ã¯ã€TDA ã€Œãƒˆãƒãƒã‚¸ã‚«ãƒ«ãƒ»ãƒ‡ãƒ¼ã‚¿ãƒ»ã‚¢ãƒŠãƒªã‚·ã‚¹ã€ (Topological data analysis) ãŒæ¥ã‚‹ ? ï½ž ãã®æ¦‚è¦ã¨ã€Rè¨€èªž / Pythonè¨€èªž å®Ÿè£…ãƒ©ã‚¤ãƒ–ãƒ©ãƒª ã‚’ã¡ã‚‰ã£ã¨èª¿ã¹ã¦ã¿ãŸ - Qiita
äººå·¥çŸ¥èƒ½ã‚„ãƒ‡ãƒ¼ã‚¿è§£æžã®é ˜åŸŸã§ã€è†¨å¤§ãªé‡ã®ãƒ‡ãƒ¼ã‚¿ã‹ã‚‰ã€ç‰©ä½“æ¦‚å¿µï¼ˆç‰©ä½“è˜åˆ¥ï¼‰ã‚„éŸ³å£°ãƒ»ä½“ã®é‹å‹•ã‚¹ã‚ãƒ¼ãƒžæ¦‚å¿µã‚’è‡ªå‹•çš„ãƒ»è‡ªå¾‹çš„ã«ç²å¾—ã§ãã‚‹å¯èƒ½æ€§ã‚’ã‚‚ã£ãŸæ‰‹æ³•ã¨ã—ã¦ã€Deep Learningï¼ˆãƒ‡ã‚£ãƒ¼ãƒ—ãƒ»ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°ã€ã€Œæ·±å±¤å¦ç¿’ã€ï¼‰ã¨ç·ç§°ã•ã‚Œã‚‹æ‰‹æ³•ãŒæ³¨ç›®ã‚’é›†ã‚ã¦ã„ã¾ã™ã€‚ ã“ã†ã—ãŸä¸ã€ä¸€éƒ¨ã®ç ”ç©¶è€…ã‚„ãƒ™ãƒ³ãƒãƒ£ãƒ¼ä¼æ¥ã®çµŒå–¶è€…ãƒ»ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã®ä¸ã«ã¯ã€ã“ã®Deep Learningã®ã•ã‚‰ã«å…ˆã‚’ã„ããƒ‡ãƒ¼ã‚¿è§£æžæ‰‹æ³•ã¨ã—ã¦ã€TDAï¼šãƒˆãƒãƒã‚¸ã‚«ãƒ«ãƒ»ãƒ‡ãƒ¼ã‚¿ãƒ»ã‚¢ãƒŠãƒªã‚·ã‚¹ (Topological data analysis) ã¨å‘¼ã°ã‚Œã‚‹æ‰‹æ³•ã«ç€ç›®ã—ã¦ã„ã‚‹æ–¹ãŸã¡ãŒã„ã‚‹ã‚ˆã†ã§ã™ã€‚ Deep Learningã¨åŒã˜ãã€äººé–“ãŒè„³ã‚’ä½¿ã£ã¦ã€äº”æ„Ÿã§å¾—ãŸå‘¨å›²ã‚„è‡ªåˆ†ã®ä½“ã«ã¤ã„ã¦ã®æ„Ÿè¦šæƒ…å ±ã‹ã‚‰ã€ã©ã®ã‚ˆã†ã«ã—ã¦å‘¨å›²ã®ç’°å¢ƒç©ºé–“ã«å¯¾ã™ã‚‹èªè˜ã‚¤ãƒ¡ãƒ¼ã‚¸ã‚„ã€ç‰©ä½“æ¦‚å¿µã€ã¿ãšã‹ã‚‰ã®èº«ä½“æ¦‚å¿µã‚’å¾—ã¦ã„ã‚‹ã®ã‹ï¼Ÿã¨ã„ã†ã€Œå•ã„ã€ã«å¯¾ã—ã¦ã€çµ±è¨ˆå¦ã‚„ç¢ºçŽ‡è«–ã®ã‚¢ãƒ—ãƒãƒ¼ãƒã§ã¯ãª
p_tan 2016/10/08
æ©Ÿæ¢°å¦ç¿’

æ•™å¸«ãªã—å¦ç¿’

topological data analysis

TDA
ãƒªãƒ³ã‚¯
1