Can you identify this sighting?
Identification history
Comments from moderator
I'll reject this sighting, rather than delete it because of the comments.| Insufficient or inconclusive evidence | 1 Mar 2023 | MichaelBedingfield | ||
| Harpobittacus australis | 15 Feb 2023 | CarbonAI | ||
| Unidentified | 15 Feb 2023 | MichaelBedingfield |
Identify this sighting
Please Login or Register to identify this sighting.
User's notes
TESTING
5 comments
MichaelBedingfield
wrote:
15 Feb 2023
Hi @AaronClausen ... this is the test I was thinking about for the CarbonAI, with an insect with long dangly legs. The photo is of a crane fly, Leptotarsus sp.(genus), which is in the category of Crane flies, midges, mosquitoes & gnats:- Other Shrubs However, since I have chosen the category of Scorpionflies and Hangingflies, it seems to only check within that category for a match. There are no crane fly species with over 100 photos. Would that make a difference if Crane flies etc were switched on and over 100 photos?
AaronClausen
wrote:
15 Feb 2023
Hi @michaelb - you are spot on in that @CarbonAI currently makes predictions only against species within the currently suggested category that have been trained (I.e. have at least 100 images). So in this instance, it had no way of knowing to jump over and look into another sibling category like Crane flies, midges, mosquitoes and gnats etc.
In the future we will enhance it to go further and automatically determine the correct category for a sighting first and then look for species predictions after that. This would also allow uses to benefit from machine learning for sightings which only have very broad suggestions like "Plant" etc, because the robot will determine that it's "some kind of daisy" before trying to predict which actual daisy it might be within its machine learning model.
In the future we will enhance it to go further and automatically determine the correct category for a sighting first and then look for species predictions after that. This would also allow uses to benefit from machine learning for sightings which only have very broad suggestions like "Plant" etc, because the robot will determine that it's "some kind of daisy" before trying to predict which actual daisy it might be within its machine learning model.
MichaelBedingfield
wrote:
15 Feb 2023
Sounds very good to me Aaron. That would be very helpful especially for insects. A lot of newer members just choose the insect category & not the sub-category, eg Moths. So the moderators don't get an email alert, so may not act on it etc.
CarbonAI
wrote:
18 Feb 2023
It's a pleasure doing this job @michaelb - remember if you don't agree with any of my suggestions, please help me improve by confirming the correct species.
Sighting information
- At least 1 Abundance
- 15 Feb 2023 11:27 AM Recorded on
- MichaelBedingfield Recorded by
-
Collections
MichaelBedingfield Rejected
Additional information
- 25mm to 50mm Animal size
Record quality
- Images or audio
- More than one media file
- Verified by an expert moderator
- Nearby sighting(s) of same species
- GPS evidence of location
- Description
- Additional attributes
