Can Popper performs incrementally learning?

[Zhang-Liao]

Dear developers,

I am trying to find a state-of-the-art ILP system for incremental learning. As far as I know, Popper is a state-of-the-art system. But I cannot understand whether it can perform incremental learning from the manual.

If it cannot, would you like to suggest me another state-of-the-art ILP system for incremental learning?

[andrewcropper]

What precisely do you mean by incremental learning? Learning different tasks incrementally?

…

On 14 Dec 2022, at 08:57, Zhang-Liao ***@***.***> wrote: Dear developers, I am trying to find a state-of-the-art ILP system for incremental learning. As far as I know, Popper is a state-of-the-art system. But I cannot understand whether it can perform incremental learning from the manual. If it cannot, would you like to suggest me another state-of-the-art ILP system for incremental learning? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Zhang-Liao]

I am sorry for my unclearness. There should be only one task.
I think incremental learning is equal to life-long learning.
For a sequence of examples E1, E2,...En. We input examples one by one. First, learn a model for E1. If we input E2, then update the original model. We can repeatedly input a new example and update the original model.
That is what I mean. As far as I know, ILP is suitable for such usage.

[andrewcropper]

Ah yes, I understand. That feature is not currently implemented in Popper but should be straightforward to implement. We will soon release a new version of Popper. We could try to include that feature in the release. Can you email me at ***@***.*** ? It might be easier to discuss this feature via email.

…

On 14 Dec 2022, at 10:22, Zhang-Liao ***@***.***> wrote: There should be only one task — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>