> Although that was only implied without any technical discussion as to why the distrust.
Good point. Computer vision systems are very fickle wrt pixel changes and from my experience trying to make them robust to changes in lighting, shadows or adversarial inputs, very hard to deploy in production systems. Essentially, you need tight control over the environment so that you can minimize out of distribution images and even then it’s good to have a supervising human.
If you’re interesting in reading more about this, I recommend looking up: domain adaptation, open set recognition, adversarial machine learning.
The discussion is missing the point of the original snarky comment
So you don't trust the computer vision algorithm...
But you do trust the meatbags?
Reminds me of the whole discussion around self driving cars. About how people wanted perfection, both in executing how cars move and ethics. While they drove around humans every day just fine
somehow it seems not as magic as setting the mathematically verifiable acceptance criteria that fails 99% of the time. (percentage chosen to show absurdity of claiming that mathematically verifiable acceptance criteria is inherently superior)
no I don't think humans are trustworthy, I think the procedures discussed are more secure than the alternative on offer which an expert in that technology described as being untrustworthy, implying that it was less trustworthy than the processes it was offered as an alternative to, and then gave technical reasons why which basically boiled down to the reasons why I expected that alternative would be untrustworthy
Good point. Computer vision systems are very fickle wrt pixel changes and from my experience trying to make them robust to changes in lighting, shadows or adversarial inputs, very hard to deploy in production systems. Essentially, you need tight control over the environment so that you can minimize out of distribution images and even then it’s good to have a supervising human.
If you’re interesting in reading more about this, I recommend looking up: domain adaptation, open set recognition, adversarial machine learning.