alt Hacker News> We did in fact need this to get AI to recognise cats.
I believe that this is simply because of the way we train ML, with labelled data. It is quite conceivable that we could get an ML model to recognise cats just by some form of multidimensional clustering of training data.
Replies
the quote from the article is just a contrived tautology that misunderstands the nature of the problem. the dualism problem is not about finding an explanation for why what we call a cat is what we call a cat. it's that you can measure anything you want but nevertheless despite confidently establishing the size of a cat, the appearance of a cat, the behavior of cats, a sophisticated taxonomy of related species labeled "felines", surveying people to find out what cats are to them, what a cat is to me is not what a cat is to you
I wish I'd phrased it better, my point was more that early vision systems had weird issues, which we were able to figure out by looking at what part of the image those models paid attention to and realising it often wasn't even part of the animal in the photo, but e.g. the plants around them. We literally had to think about what made a cat a cat to make AI good at recognising cats.
This would also impact clustering.
That said, I think even for humans there's a similar issue: we spent millennia clustering things into groups and labelling those groups, which is why the Catholic church had rules about no meat on Good Friday but fish was fine and beavers counted as fish (and there is now a podcast titled around the idea there is no such thing as a fish*). For cats, I don't see it myself but the fossa is described as "cat-like".
* https://en.wikipedia.org/wiki/No_Such_Thing_as_a_Fish#Title