Neumann's Starling as a Suggested ID

[kylayuza]

I’ve been spending a lot of time lately going through a backlog of unconfirmed bird nest observations in the eastern United States, and I’ve noticed that Neumann’s Starling (Onychognathus neumanni) - https://www.inaturalist.org/taxa/145112-Onychognathus-neumanni - gets suggested as an ID for photos of bird nests that look like clumpy masses of grass. There are a few issues with this:

• Neumann’s Starling is a west African species never reported in North America
• There are less than 15 observations of this species on iNat, none of which are of nests
• Apparently there’s an issue with users removing the suggested species ID from observations before they’re submitted (https://www.inaturalist.org/observations/167579930)

I know that ID algorithms aren’t perfect and that bird nests in particular are troublesome, but I’m confused about this one since the photos aren’t visually similar and aren’t remotely close to where the species is found normally. I’m new to the iNat forum so I don’t know if this is a common issue, a bug, etc., but has anyone else encountered this sort of problem before and is there something that can/should be done about it?

[spiphany]

If you look at the list of names on the taxon page, it seems that NEST has been entered as an English name for Onychognathus neumanni. So when people type in “nest” (because they saw a bird’s nest) the database finds this species.

You might create a flag (under the link “curation” on the taxon page) and outline the issue with the common name and see if it can be edited/removed.

[dianastuder]

Those 4 letter bird abbreviations are very confusing - for the rest of us who don’t use them as a default. I tripped over one before. Which was removed, rolled down the list instead of being FIRST. They need to be in the list for birders who do use them.

I’m not a birder, but that common name - NEST - shouldn’t be listed as ‘English’ it should be ‘whatever it is Code’
https://www.carolinabirdclub.org/bandcodes.html

[thebeachcomber]

just to slightly expand on @spiphany’s great answer, this issue has nothing to do with the computer vision; the species is not in the model, so it’s impossible for it to be suggested. Any erroneous ID you see of this species has been suggested by a person manually typing in the word ‘nest’ as explained above.

To tell if an ID comes from a computer vision suggestion or not, look out for this symbol

[kylayuza]

I had no idea! That’s good to know, thank you!

[dianastuder]

But remember - lots of identifiers use CV as a shortcut.
CV is a tool, like a field guide, or a taxon-based website.

[srushtibokil]

Here in India it doesn’t only show up on nests- even red ant houses
Really annoying to see

[kylayuza]

Thank you for the thoughts, everyone. I flagged Neumann’s Starling the other day and the taxon has since been updated, so this issue shouldn’t occur anymore.