Looks as if there are no boundaries when it comes to AI. @Janice_A@miked@Luisa
Luisa and I have been interacting on iSpot ZA and I asked if she could help me with a caterpillar from the Cape West Coast - a long way from Europe!!
Our interaction started with a Bird ID and then on to this Caterpillar
The name Luisa found on SEEK had been observed on iSpot before the Great Trek by some folk - so I assume the obs. were moved across the Atlantic.
Whatever - I followed up on Wikipedia and confirmed the link to the the name of the moth I’d posted that was agreed here
Think you’ll like this @dejayM - thanks for your support always
Team work helps - t think very exciting to see AI can also work
Remarkable if AI can deal with a variable species that can actually change appearance over time and there are loads of other similar catterpillars. The basic question is of course how many of the possible species are there in the AI, suspect only a tiny number of the total possible species that could look like this. On the other hand it is likely to have the most common species in, so long as it is a relatively common species then may well be identified correctly.
A question is whether the identificatoins will be revised each time the AI is updated with a fuller range of species and its model improves. This has happened for the UK mammals on iSpot, the AI updated its training set and had another go(on a test not the live site), that is relatively easy for the AI but not for the humans who have to go back and check what it is now saying is any better (or worse) than previously.
Impressive comments record about what seems to be a common species, with a ‘who’d a thought it’ life cycle. But as some comments say, common does not necessarily mean well understood.
