Tiliqua rugosa

2 Shingleback Lizard at Ainslie, ACT

Request use of media

Identification history

Tiliqua rugosa	30 Jan 2023	WillO
Tiliqua rugosa	30 Jan 2023	CarbonAI
Unidentified	30 Jan 2023	AaronClausen

Identify this sighting

Please Login or Register to identify this sighting.

9 comments

AaronClausen wrote:

30 Jan 2023

Wow thanks for the quick ID @CarbonAI - we've been expecting you..

AaronClausen wrote:

30 Jan 2023

@WillO thanks for the confirmation. You just help reinforced and train the suggestion made by our new AI non-human moderator - Carbon AI. When you as an expert human moderator confirms what Carbon AI suggested, it reinforces it's machine learning model that it was correct and if you disagree with it, that also helps it learn and get smarter over time. Key goal is to alleviate the workload of the more common/mundane species of our expert human moderators so human experts can focus on the cool/tricky/difficult stuff.

RodDeb wrote:

30 Jan 2023

@CarbonAI wow that is very clever @AaronClausen

AaronClausen wrote:

30 Jan 2023

Thanks @RodDeb has been in the works for a while. We just want to try to ensure we protect and don't wear out our overloaded moderators because they get a lot of mundane sightings to confirm e.g. mapgies or kangaroos etc. A robot can easily be trained to do those with guidance from the human experts. But the robot will struggle with anything tricky and that is exactly where we want the precious time of our expert moderators to be best invested. The robot basically works for the community and we can all train it over time to be more useful.

RodDeb wrote:

30 Jan 2023

@AaronClausen will be wonderful if it works on the more common flora and fauna to lighten the load a bit for all the hardworking moderators. We really appreciate all the hard work they do so anything to help them is great. So clever.

AaronClausen wrote:

30 Jan 2023

Spot on @RodDeb . We'll gradually expand the number of categories that it will monitor to hopefully include those. Currently only doing kangaroos and skinks while we bed it in.

jks wrote:

30 Jan 2023

So good @AaronClausen

MatthewFrawley wrote:

30 Jan 2023

This is very exciting. I reckon @CarbonAI would be good at correctly identifying the Meadow Argus and Australian Painted Lady. Could it also learn to ID the Common Brown with the male and female looking different? I'm guessing it could when you see what AlphaZero could accomplish.

AaronClausen wrote:

31 Jan 2023

For sure @MatthewFrawley great questions, let's give it a crack. I'll setup training for butterflies today. It just finished learning daisies last night.

Note: to trigger Carbon AI at the moment, a sighting needs to be IDd to the bottom category level e.g. Nymphs. Otherwise it won't currently get triggered.

Please Login or Register to comment.