Compare tool has no results sometimes

I’ve been on iNat for a few months and am still puzzled by why the Compare tool shows no suggestions sometimes.Or it only shows one suggestion when I know there a numerous genera or spp that could qualify. (This is after there is at least one suggestion, of ten made by the original poster.)

Is it because this sp. has not been seen near this location? Is it because it is impossible to ID the specimen to this level without special qualifications that iNat knows I don’t have?

(I tried to find a previous discussion on this but had no luck.)

It took me a while to figure out how the compare tool works. Basically, if there is already a species ID, the Taxon option will default to one level higher in the taxonomic ladder, you might want to change that.

Screenshot from 2020-06-14 22-45-04

Moreover, the default source is observations, which indeed only shows taxa that have been observed in the specified place, you might want to select a different place (for example going from municipality to region or higher) if you don’t have any options to select from.

If you want to use the computer vision, you have to switch source to the visually similar option. Another source that could be useful if you suspect the identifier made a mistake is misidentifications (I believe it shows which taxa are usually IDed as that species, and then corrected), but I’ve never had much luck with it personally (you might want to select a broader option for place here as well, or remove it completely, it won’t have enough data if the place is small). The checklist option can also be useful, if the place you are looking at has one.


Thanks for that. So it’s not perfect! Or I’m not using it perfectly! :)

I am willing to bet that others like me are very tempted to select the ID chosen by the system even if you’re not personally sure. I have started putting a note that the ID was selected by iNat and not due to my knowledge (I now know there is a little marker for when that is done, which I assume allows researchers to flag IDs that were not improved upon or confirmed by third parties).

1 Like

Ps. Is computer vision the same as the observations system? Or are there two A.I. systems?

I’m not sure I understand what you mean.

When I say computer vision I mean using the image recognition. I have noticed that it tends to give different results if done from the app, as opposed to using the compare tool. Not sure if that’s what you intended?

Thanks. That is what I meant, but I did not notice that about the different results. That’s interesting too.

I think this whole iNat thing is a bit puzzling. It makes sense for birds and mammals but when it comes to small things or species, it seems misleading to present suggested IDs to people when those people really can’t be sure those suggestions are correct. I got into the habit for a few days earlier on of just assuming the system was correct if it presented me with one suggestion when I was a the genus level, but I’ve stopped doing that because I’m really not an expert, and I know enough about entomology to know that it’s usually invisible characteristics that make the difference.

No need to reply, lvdt. I’m just ruminating about how all this works and what I can contribute accurately.


This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.