Flag for Taxon: Long-tailed Aphideater (Eupeodes americanus)
| Flagger | Content Author | Content | Reason | Flag Created | Resolved by | Resolution |
|---|---|---|---|---|---|---|
| jeremyhussell | Long-tailed Aphideater (Eupeodes americanus) |
iNat computer vision suggests this species way, way too much |
Oct. 29, 2019 11:57:15 +0000 | bouteloua |
maintenance mode now |
Comments
Yeah, it definitely does, it also suggests the genus Syrphus way too much for observations that are actually Metasyrphus sp., probably 2/3 of the observations of Syrphus are actually of Metasyrphus.
Just went through them all, so now there are 7 observations instead of ~350.
oof. thanks for all your work!!
I think the number was pretty stable before, and now all the people who were identifying them know the situation, so I think we should be good now...
Thanks @upupa-epops!
@bouteloua suggested flags as the best place to coordinate cleaning up species with a lot of misidentifications due to overconfident computer vision. See https://forum.inaturalist.org/t/computer-vision-clean-up-wiki/7281
There are two issues here:
1) The computer suggests this species for many Syrphini, e.g. Syrphus, Epistrophe, Parasyrphus, etc.
2) E. americanus and E. pomus are nearly indistinguishable, but due to E. americanus being more common and a feedback loop on iNat, there are 0 observations of E. pomus and a few hundred of E. americanus.
I think many E. americanus observations should be knocked back to tribe Syrphini, and most of the rest should be at subgenus Metasyrphus.