Complexes potentially not listed as ID suggestions

Bug Reports
1

Platform (Android, iOS, Website): Website/general computer ID suggestions

URLs (aka web addresses) of any relevant observations or pages: https://www.inaturalist.org/observations/58383587

Screenshots of what you are seeing:

iNat
iNat696×718 113 KB

Description of problem:
Step 1: Submit ID
Step 2: Get ID suggestions
Step 3: Complex missing from the list

Hi, I am not sure if this is just a one time thing, known issue, or otherwise but I figured it was worth bringing to attention that when submitting my millipede ID the complex for which the observation belongs was not provided under ID suggestions. Since these are not field identifiable (as is the case many bugs and other organisms) this likely leads to a lot of casual observers choosing to ID to a species that is not field identifiable. Admittedly I was guilty of this as this is the species I thought it was from the info that I had (I know, I know, I should have went with genus).

Probably wishful thinking but would be really cool if species within non-field identifiable complexes were removed from the suggestions (but left in the database of course for those who put things under a microscope or otherwise for ID) and replaced with only the complexes themselves. Though that’d probably require some pretty high tech computer wizardry!

Best,

Nate

5 Likes
2

Well, https://www.inaturalist.org/taxa/904631-Narceus-americanus should be in the model, it’s one of 15 species complexes exported when the model was trained last year. As to why it wasn’t coming up as a suggestion in this particular case, we’re not sure at the moment. Has anyone seen a complex appear as a computer vision suggestion?

3

I looked through a few complex lists and thumbed around different observations within them, and out of about 10 or so across 4 species I was able to get one to populate. The others did not include the complex in the suggestions list. After seeing this one I tried a few others in the European bramble complex and it pretty consistently came up as a suggestion with observations that had reasonable photos.

iNat2