This topic is to begin discussion on developing the feature of automated Feature selection. I would like to start of first for only time series data.
I think the main problem with this is the fact that at the moment there is a 1:1 mapping between input columns and placeholders in tf, if one adds some process that would map from one input to potentially several inputs, that would require some work.
Anyway, can you please clarify exactly what is your use case and how do you imagine the process to look like? Thanks.