Community ID: Disagreements with previous ID are interpreted as disagreements with future IDs of totally different species

Bug Reports
1

See this observation: https://www.inaturalist.org/observations/7796113

Since the IDs of this observation will probably change following this post, here are the IDs:

Grasshopper
Typical Cicadas
Typical Cicadas
Typical Cicadas
Resh Cicada
Resh Cicada

The community ID: Typical Cicadas! The problem is the algorithm interprets Typical Cicadas as a disagreement with the Resh ID. I may misunderstand this, but it will now take five more Resh IDs to get the community ID to Resh. This is a HUGE problem, because this situation happens all the time! The reason for for the three Typical Cicadas IDs is that the observation won’t even show up as a Cicada until there are three of them to get the community ID to cicada (so the cicada people will even know it exists).

If the IDs were instead these:
Typical Cicadas
Typical Cicadas
Typical Cicadas
Resh Cicada
Resh Cicada

The community ID would be Resh because in this case, Resh is interpreted as improving on the Typical Cicada ID, not conflicting with it.

1 Like
2

No it doesn’t. The Grasshopper ID is the only one disagreeing with Resh Cicada.

image
image1146Ă—662 80.7 KB

Because Resh Cicada has a score of .667, you only need one more to get over the 2/3 requirement.

3 Likes
4

Then the bug is that the grasshopper ID disagrees with Resh. If there have been enough “Typical Cicada” ids to confirm that it’s not a grasshopper, the grasshopper ID should have no effect the ability to refine the cicada identification.