Generally speaking, most Notified Bodies don't classify a machine learning model update as a significant change if your performance improves and your model architecture remains the same. You need to re-do your verification of your new model (e.g. do some testing with a test set and document the results). Active learning models which change their weights during deployment are currently not allowed.

Regulation and notified bodies are only slowly catching up with machine learning - based software. There’s no official standard or guidance so far (is that a good thing?) so my answer is entirely based on my experience with auditors.

An update of your machine learning model, i.e. an update of its weights, is generally not seen as a significant change. You need to re-do the verification of your ML model and store it internally.

This makes sense, because your ML model architecture stays the same and the performance has probably improved - so there probably aren't huge risks involved.

As there's no official guidance on this yet, you'll have to confirm with your auditor whether this process works with them.

If you change the architecture of your ML model, you may face more scrutiny but just be sure to prove that the decision made sense, your performance hasn’t gotten worse and you haven’t made the risk profile of your software worse.