Should we call deep learning “statistics”? It’s mostly empirical. Should recommendation systems be called “statistics”? What about multi objective non convex optimization?

Not everyone is doing regression and classification all day.

The above were studied in the Computer Science curriculum at my university. I work with a statistician, like someone with a degree in mathematical statistics. They know nothing about any of those.

> Not everyone is doing regression and classification all day.

Yeah, some of them are doing unsupervised learning (recommender systems) too!

I dunno, I personally think we'd all have been better off if we'd called it statistics as at least then people would realise that the field wasn't created yesterday.