Need help with reID

Yes, something like this is part of what I was trying to suggest. If the goal is to recruit users who are not experts in the taxon to help re-ID observations, it is helpful to provide them with some basic guidance about the relevant characteristics are for distinguishing the possibilities within the group in question. Expecting people to blindly push back IDs to genus level based merely on range without knowing anything about how to ID them does not seem like a good idea for multiple reasons.

From a language standpoint, I am happy to consult extant material in German, but this is likely not the case for everyone, and non-expert users may struggle to use a scientific key for a taxon that is difficult to distinguish in the first place. So creating a guide yourself – while a lot of work – may be a better way to reach out to these users.

These tips do not necessarily need to be as detailled as a professional key – if, say, a species ID is not possible without seeing what kind of hair is on the stem, people can use this information to at weed out those observations where the stems are not visible. Or if all European Thymus species with a certain feature are in a specific section or subsection, users can provide some general sorting even if they aren’t skilled enough to make finer IDs.

On the other hand, if you hope to recruit botanists who already have expertise with Thymus, the forum may not be the most effective way to get their attention.

4 Likes

The Flore de la France méditerranéenne continentale confirms this and lists T. serpyllum as “douteuse” in France and usually amalgamated with the section serpyllum.
The key itself is clear too: either T. vulgaris or one in the section serpyllum which is “d’un abord très difficile”.

So, mea culpa, I didn’t read that and relied on the iNaturalist machine.

The Guide de la Flore 06 lists just two Thymus: T. vulgaris and T. serpyllum but notes that they include T. praecox, pulegoides, polytrichus and more in the latter. So they actually mean the section serpyllum, not the species T. serpyllum.
So they oversimplified the guide in this regard.

I’m curious if the machine learning algorithm is going to adapt to the corrections of the T. serpyllum IDs.

1 Like

CV is updated once a month roughly. Definitely worth the effort to improve the data it uses.
https://www.inaturalist.org/blog/87875-our-first-new-computer-vision-model-v2-10-for-2024-including-1-599-new-taxa

4 Likes

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.