Continual Learning in Tabular / Wide Data

Algorithms
#1

Do you know works/models dealing with unconventional CL data formats as input, such as tabular and/or wide data using (deep) neural networks, where CL (or regular deep representation learning) strategies may be helpful for classification?

(*) wide in this case means tabular data with larger amount of columns (i.e., features) than rows (i.e. data points).

1 Like
#2

Hi @natalia! I didn’t quite get your questions! to what data types are you referring to specifically? :slight_smile:

#3

I re-edited my question.

1 Like
#4

Mmh, that’s interesting, maybe this can be seen as a few-shot setting?

#5

Yes, but here we may not be facing previously-unseen classes.

1 Like