When to use random forest vs deep learning for structured data?

I am going through ML1 and DL1 and trying to get a grasp on what problems will be best suited for using random forests vs deep learning. Is it getting to the point where deep learning can out perform random forests for most structured data problems? Is one approach better for regression/classification?

With my current project, I can’t quite get the deep learning approach to beat using a random forsest, but could definitely be doing something wrong.

Anyone with experience testing each approach out on the same data set would be helpful. Thanks :slight_smile: