
ç»åçæAIã®æã«é¢ããç ç©¶ #1 - äœæ AIã¯æãèŠæãªã®ã
ã¯ããã«
ç»åçæAIã§ã¯ãæã®æç»ã«å€§å€èŠæ
®ããå Žé¢ãå€ãèŠåããããŸãã
NegativeTIãLoRAãLyCORISãªã©ã§æã«é¢ããŠã®æ¹åæ¡ãæ¥ã
ç ç©¶ãããŠããŸãã
ã§ã¯äœæ
ãããããAIã¯æã®æç»ãèŠæãªã®ããçŽè§£ãã€ã€ãå¥ã®ãã¯ãã«ããæ¹åã§ããããæš¡çŽ¢ããŠãããŸãããã
ãŸãããã®èšäºã¯ã·ãªãŒãºåããäºå®ã§ããæ¬èšäºã§ã¯å®çµããŸããã
äœæ ãAIã¯æãèŠæãªã®ã
AIãšäººéã®æèåè·¯
ãããçŽè§£ãã®ã¯ãStable DiffusionããŸãã¯Latent Diffusionã®åŠç¿æŠå¿µãšãã€ãºç»åãã埩å·åããããã«äœ¿çšãããU-Netã®åŠçã絡ãã§ããŸãã
U-Netã«ã€ããŠã¯ã以åéå±€ããŒãžã®æã«èª¬æããŸããã

Fig. 1. U-net architecture (example for 32x32 pixels in the lowest resolution). Each blue
box corresponds to a multi-channel feature map. The number of channels is denoted
on top of the box. The x-y-size is provided at the lower left edge of the box. White
boxes represent copied feature maps. The arrows denote the different operations.
as input. First, this network can localize. Secondly, the training data in terms
of patches is much larger than the number of training images. The resulting
network won the EM segmentation challenge at ISBI 2012 by a large margin.
Obviously, the strategy in Ciresan et al. [1] has two drawbacks. First, it
is quite slow because the network must be run separately for each patch, and
there is a lot of redundancy due to overlapping patches. Secondly, there is a
trade-off between localization accuracy and the use of context. Larger patches
require more max-pooling layers that reduce the localization accuracy, while
small patches allow the network to see only little context. More recent approaches
[11,4] proposed a classifier output that takes into account the features from
multiple layers. Good localization and the use of context are possible at the
same time.
In this paper, we build upon a more elegant architecture, the so-called âfully
convolutional networkâ [9]. We modify and extend this architecture such that it
works with very few training images and yields more precise segmentations; see
Figure 1. The main idea in [9] is to supplement a usual contracting network by
successive layers, where pooling operators are replaced by upsampling operators.
Hence, these layers increase the resolution of the output. In order to localize, high
resolution features from the contracting path are combined with the upsampled
https://arxiv.org/pdf/1505.04597.pdf
U-Netãšã¯ãã€ãºç»åããçæç»åãæšè«ããéã«ãç©äœã®å¢çïŒã»ã°ã¡ã³ããŒã·ã§ã³ïŒãæšå®ããçºã®ãããã¯ãŒã¯ã§ãã

Stable DiffusionããŸãã¯Latent Diffusionã®åŠç¿ææ³ã¯ä»¥äžã®éãã§ãã

https://github.com/CompVis/latent-diffusion
ãã®ããã«ãæ¡æ£ã¢ãã«ãšã¯ãã€ãºç»åããã»ã°ã¡ã³ããæšè«ããŠçæããŠããŸãã
ããŒã¿ã»ããåŠç¿ã¯ãã®éã§ãç»åãäžåºŠãã€ãºåãããããå
ç»åã«åŸ©å
ãããšãã£ã圢ã§åŠç¿ããŠããŸãã
ããå€ãã®æ¹ãééããããã®ã§ãããAIã¯ãã®ç©äœãäœãªã®ãããã®æŠå¿µãèªèããŠããããã§ã¯ãããŸããã
èšããªãã°ã人éã®æã¯åå5æ¬ã§ããããšæã 人éã®æŠå¿µãåŠç¿ããŠããããã§ã¯ãªããšããããšã§ãã
ãŒããããšæ³åããŠã¿ãŠãã ããã
ãã®ãã€ãºç»åã«äººéã®ãããªæŠå¿µãååšããã§ããããã
AIåŠç¿ãšèããšã人éã®ãããªåŠç¿ãã€ã¡ãŒãžããããšæããŸãããè¿ããããã§å®éã¯éããŸãã
ããã1ã€ãã€çŽè§£ããŠãããŸãã
人工ãã¥ãŒãã³
人éã®æèãªã©ã¯ãè³å ã®ç¥çµçްèïŒãã¥ãŒãã³ïŒã«ãã£ãŠäŒéãããŠããŸãã

https://ja.wikipedia.org/wiki/%E7%A5%9E%E7%B5%8C%E7%B4%B0%E8%83%9E
ã·ããã¹ãšåŒã°ãã黿°ä¿¡å·ã®éåãè¡ã£ãŠããã¥ãŒãã³å士ã®äŒéãè¡ã£ãŠããŸãã
ãã®ä»çµã¿ãAIãžå¿çšããã®ããã¥ãŒã©ã«ãããã¯ãŒã¯ã§ãã
ãã¥ãŒã©ã«ãããã¯ãŒã¯ãšã¯ã人éã®è³ç¥çµã®ä»çµã¿ãæ°çã¢ãã«åãããã®ã§ã人éã®ãã¥ãŒãã³ãç䌌çã«æ§æãããããã¯ãŒã¯ã§ãã

äžå³ã¯åçŽããŒã»ãããã³ãšåŒã°ããã¹ããã颿°ïŒç©åœ¢æ³¢æ§ã®ã°ã©ãïŒã«ãªããŸãã
ããã¯ããã¥ãŒãã³ãåºç€çã«æ°çã¢ãã«åãããã®ã§ãããå®éã®ãã¥ãŒã©ã«ãããã¯ãŒã¯ã¯ã·ã°ã¢ã€ã颿°ïŒã¹ããã颿°ãé£ç¶æ®µéçã«è¡šçŸããææ°é¢æ°æ²ç·æ§ã®ã°ã©ãïŒãçšããŠãããšããã®ããå°ã
éã£ãŠããŸãã
ããã€ããã¥ãŒã©ã«ãããã¯ãŒã¯ã«çš®é¡ã¯ãããã®ã®ãç»åçæç³»ã§äœ¿ãããŠãããã¥ãŒã©ã«ãããã¯ãŒã¯ã¯ç³ã¿èŸŒã¿ãã¥ãŒã©ã«ãããã¯ãŒã¯ïŒConvolutional Neural NetworkïŒCNNïŒãå©çšãããŠããŸãã
ç³ã¿èŸŒã¿ãã¥ãŒã©ã«ãããã¯ãŒã¯ãšã¯ãã»ã°ã¡ã³ãæ¯ã«ç»åãèŠããšãã«ãç»åã®ç¹åŸŽãéç«ãããç¹åŸŽãæœåºããŠãããŸãã
ããã«ãã£ãŠãç»åã®ç¹åŸŽããããåããäœæããã®ããã®ç³ã¿èŸŒã¿ãã¥ãŒã©ã«ãããã¯ãŒã¯ã§ãã
ãã¥ãŒã©ã«ãããã¯ãŒã¯ã§ã¯æ§ã
ãªãã€ã¢ã¹ãæ°å€åããŠåŠçããããæ
ã«ããã®ãã€ã¢ã¹ãäœãæããŠããããããã¯ã©ãããæŠå¿µãªã®ããã©ãããææ
ãªã®ãã¯ç€ºãããŸããã
ãããŸã§ããæå³ãããã®ãã©ã®ãããªç¹åŸŽãããã®ãããæ°åŠçåŠçããŠããã«ããéããŸããã
ããã«é¢ããŠèšããšã人éã®æèã§ã¯ãã£ãšè€éã«è§£éããŠããŸãããã
ããã人éãšAIã®å€§ããç°ãªãç¹ã§ãã
AIãæãèŠæãšããåå
ããããæ¬é¡ã§ãã
ãããŸã§èŠãŠããã°ãäœãšãªããµãã£ãšæ³åã§ããŸãããã
æãšãã圢ç¶ã¯ãå€é¢ç¯ã§ããããã€æ§ã
ãªéãªããããŸãã
AIãæãèŠæãšããåå ãšããŠæããããã®ã¯ä»¥äžã§ãã
ããããæïŒæ5æ¬ã§ãããšããæŠå¿µèªäœãAIã¯æã£ãŠããªã
æïŒæ5æ¬ãšããæŠå¿µãå šããªãç¶æ ã§ãæãšãããã®ã®ç¹åŸŽãå°åãã«ããç©äœã®ã»ã°ã¡ã³ãåäœã§ããã®å¢çãæ°å€åããç¹åŸŽãç®åºããŠããã ãã«éããªã
æ§ã ãªéãªããããæã®ã·ãã¥ãšãŒã·ã§ã³ã¯ãè€éãããŠç¹åŸŽã®æ°å€åãããéãç®åºçµæãæåŸ å€éãã«åŠçãããªã
ããããé¢ç¯ãã©ã®ããã«æ²ãããããšããå¯åç¯å²ã®æŠå¿µãAIã«ã¯ååšããªã
çªã®å æ²¢ãé¢ç¯ãæã®çºãªã©ã¯é«åºŠãªã¬ã³ããªã³ã°æè¡ãå¿ èŠã§ãããçŸæç¹ã®Generative AIã§ã¯åçŸããã®ã«é床ããã
ã€ãŸãã人éçæèã§èšãã°ããã¹ã¿ãèŠãããšãé£ã¹ãããšããªã人ïŒãã¹ã¿ãšããæŠå¿µããªã人ïŒã«ããã¹ã¿ã®ç¹åŸŽã ããäžããŠïŒåŠç¿ãããŠïŒããã¹ã¿ãäœãïŒçæããïŒããšèšã£ãŠãã®ãšåããªã®ã§ãæåŸ å€éãã®çµæã«ãªãã«ããããšããããšã§ãã
ã§ã¯ãã©ã®ããã«ã¢ãããŒãããã®ã
å®ç§ã«å¯ŸåŠããã®ã¯çŸç¶ç¡çã§ãããæã«ç¹åããåŠç¿ã¢ãã«ãããã®ãäºå®ã§ãã
ããããæã«ç¹åããåŠç¿ã¢ãã«ã ãšãåã
ãçæããããã®ãšã¯ä¹é¢ããããŠããããŸãã
ã§ã¯ãäžãããã€ã¢ã¹ã»æ å ±ãä»å ãããŠãããã°ããã®ã§ã¯ãªããïŒïŒç¹åŸŽã®æ°å€åã«å€åãä»ããïŒãšããã¢ãããŒããåãããšæããŸãã
ãšãŠãããäŸããã¶ããºãããïŒ@blue_pen5805ïŒããã€ãŒãããwakame74æ³ã§ãã
ã¡ãŸãã§åã®wakame74æ³ãç¥ã£ãŠããŸããïŒ
â ã¶ããºã/blue.pen5805 (@blue_pen5805) May 9, 2023
ããã䜿ããšã€ã©ã¹ãå šäœã®å質åäžïŒæã®ç Žç¶»çãå€§å¹ ã«äœæžã§ããŸãïŒ
æ·»ä»ã®ã€ã©ã¹ããèŠãŠããã ããã°åããããããŸãïŒïŒå šéš t2i æªä¿®æ£ïŒ
äœ¿ãæ¹ã¯ç°¡åïŒããã³ããæ«å°Ÿã«74ããŒã¯ã³åã® wakame ãæ·»ããã ãã§ãïŒè©³ããã¯ALTãèŠãŠãïŒ#AIArt pic.twitter.com/RRbndhhG1b
wakameãããŒã¯ã³åãããšã2ããŒã¯ã³äœ¿çšãããŠããŸãã

wakã11172ãameã780ãšããŠæ°å€åãããã¥ãŒã©ã«ãããã¯ãŒã¯ã«æž¡ãããŠããŸãã
ä»åã®ã¢ãããŒãã¯ãã®ããŒã¯ã³æ
å ±ãéµã«ãªããŸãã
ããã¯æ¬¡åã®ç ç©¶ã·ãªãŒãºèšäºãŸã§ãåŸ
ã¡ãã ããã
ãããã«
ä»åã¯ãåã®ã·ãªãŒãºåã§ããç ç©¶èšäºãç»åçæAIã«ãããæã«é¢ããç ç©¶ããäŒãããŸããã
Generative AIã«ãããŠãæã¯æå€§ãã€æãå°é£ãªèª²é¡ã§ãã
å®ç§ã«åçŸããããšã¯ã§ããªããããããŸããããå°ãã§ã劥åã§ããéšåãå人çã«ç ç©¶ããŠãããããšæã£ãŠããŸãã
次åã®ç ç©¶èšäºããæ¥œãã¿ã«ïŒ