Sean Brattin requested Google’s John Mueller a few months in the past if Google makes use of Cloud Vision AI inside Google Image Search. The reply was no, however John did say that Google did speak about “about doing that previously” however ended up to not do it.
This got here up in a video web optimization hangout again in February on the 7:16 mark into the video. Sean Brattin requested “So can you touch upon this? I think about that Cloud Vision has one thing to do with that, making an attempt to match similarities with machine studying to the entities. Am I heading in the right direction right here?”
John Mueller responded “I do not know the way far we might use one thing like that. I do suppose that, at the very least so far as I perceive, we have talked about doing that previously, particularly for picture search. But it is one thing the place, simply purely based mostly on the contents of the picture alone, it is generally actually laborious to find out how the relevance ought to be for a selected question.”
John then gave this instance, “So for instance, you might need, I do not know, an image of a seashore. And we might acknowledge, oh, it is a seashore. There’s water right here. Things like that. But if somebody is looking for a lodge, is an image of the seashore the related factor to point out? Or is that, I do not know, a few miles away from the lodge? It’s actually laborious to evaluate simply based mostly on the content material of the picture alone.”
John added that Google does use methods to grasp pictures however probably not from the Vision AI. John mentioned “so I think about if or after we do use form of machine studying to grasp the contents of a picture, it is one thing auxiliary to the opposite elements that now we have. So it is not that it could fully override every thing else.”
Here is the video embed with these particulars:
Forum dialogue at YouTube Community.
