Inactive taxon suggested by computer vision

Taxon 244002, Diceroprocta viridifascia sensu lato, was inactivated during a taxon split on July 27th, 2020, but as of today, August 11th, it still appears in the computer vision suggestions (example observation).

Computer vision suggestions that were split should probably be bumped to the common ancestor of the output taxa instead (in this case, the genus Diceroprocta, which is already suggested).

6 Likes

Because there are now an active an inactive version I think. When Diceroprocta virdifascia viridifascia and v. beqauerti were split, any not at subspecies became Diceroprocta and D. viridifascia was made inactive but automatically re-created by the taxon split. (I think that’s what happened).

D. viridifascia is a valid name so it needs to be an option, I think it just got doubled up.

1 Like

From staff:

I agree with @bouteloua, it would be nice to handle splits by moving up to the common ancestor rather than suggesting inactive taxa.

This taxon split happened 8 months before the CV-suggested ID.

3 Likes

This issue is still happening:
image

Interestingly when I try to pick something from the AI suggestions, the inactive one isn’t even shown as an option for me???

Still happening v2
image

Same here
image

Please include links with examples which will help us investigate. I think found this latest observation at https://www.inaturalist.org/observations/87021865. It was generated with the Seek app, which is susceptible to suggesting inactive taxa via its CV suggestions. In theory, the website and core iOS and Android apps should not suggest inactive taxa (though they could obviously suggest taxa which later become inactive).

1 Like

My bad, it’s indeed that observation

@pleary
https://www.inaturalist.org/observations/85053579
https://www.inaturalist.org/observations/66771476

Thanks for the links. Looks like both of those observations were uploaded from the Seek app. Seek’s on-device CV suggestions are a little different from what the core iNaturalist apps and website use. We’ll look into preventing Seek from suggesting inactive taxa. If anyone notices an example that didn’t originate from Seek, that would be interesting to see.

5 Likes

is there somewhere on an observation that indicates it came from Seek?

just found two more that I assume are Seek-based, but can’t see anything at either link definitively saying they are
https://www.inaturalist.org/observations/87723486
https://www.inaturalist.org/observations/87897831

1 Like

In the right column just above the DQA section

image

(those are both Seek obs)

4 Likes

cheers, that menu was collapsed for me

2 Likes

Here’s 263 observations of Alnus glutinosa (European/Common Alder) with this same issue. There was a split, and all the recent observations appear to be from Seek.
(see this flag)

One that I looked at, https://www.inaturalist.org/observations/41104250 doesn’t appear to be from Seek. " This observation was created using:

[![iNaturalist iPhone App]

You can see the GitHub issue for this here: https://github.com/inaturalist/SeekReactNative/issues/369

1 Like

Trying to figure out why CV would appear to suggest an inactive taxon. And of all taxons, Unkown!

I come across the use of this taxon from time to time but this is the first time I have seen the CV symbol. Pretty sure this was not the result of a split. The observation was made using the iphone app.


https://www.inaturalist.org/observations/106237108

Couldn’t resist adding an ID to that … so no longer Unknown.

1 Like

An Identificologist has to do what an Identificologist does. I held back from at least adding Dicot just so at least one other person would see it.

2 Likes

But yes - exceedingly odd that iNat would or could suggest Unknown!

1 Like