Gregory's Server@baldur I object to opt-out analysis of content that’s not private. It’s not clear the analysis here matches your description though; this same text would also cover analyses such as measuring what percent of Lightroom images are landscapes, which doesn’t involve taking images and including them in a training set.
|
2 comments
 @baldur I have no idea what Adobe is doing — best we can do there is ask Adobe directly. On training: after training, one can then use a trained model to make other determinations. The other determinations could be what’s covered in this text. But again, let’s ask. |
@stig AFAICT, most current machine learning approaches involve building a training set of some sort?
Also, after its history Adobe does not get the benefit of the doubt from me.