4 ms·
Hah thanks for that! https://xkcd.com/1425/ https://xkcd.com/1425/ Still, image recognition is effectively a solved problem at this point for AI. Maybe there
by zackmorris 8y ago
Hah thanks for that!
https://xkcd.com/1425/ https://xkcd.com/1425/
Still, image recognition is effectively a solved problem at this point for AI. Maybe there are still problems with notation and terminology that prevent widespread adoption?
- stevenwoo 8y agoIs there an update to this story or is it not true or some other possibility? https://www.wired.com/story/when-it-comes-to-gorillas-google-photos-remains-blind/ https://www.wired.com/story/when-it-comes-to-gorillas-google... If you prefer a story not on wired (sometimes it shows a subscription nag for me) - https://www.theverge.com/2018/1/12/16882408/google-racist-gorillas-photo-recognition-algorithm-ai https://www.theverge.com/2018/1/12/16882408/google-racist-go...
- PhasmaFelis 8y agoI don't think that's changed, no. My hypothesis, incidentally, is that it's really hard for an AI to tell the difference between human and ape facial features at all; they're too similar. It seems easy to us because a big chunk of the human brain is devoted to face processing. The AI pretty much just knew "primate face," and knew that pale primate faces were usually human, but with a dark primate face it could only guess. (Which doesn't excuse Google for releasing it in that state, of course.)
- bluGill 8y agoThere are many different weeds, and some look a lot like non-weeds. Weeds of the same species often look different in different area. Companies are working on the problem, but you can't train on weeds in one area and go to the next.