CNN in High Dimensional Spaces / Non-CV Applications

Two questions on the use of CNN:

  1. Are CNN’s used for non-CV applications e.g. structured data with multiple features?
  2. If yes, have they been used to predict a continuous value as opposed to classification problems covered in the course?
1 Like

What do you mean by “non-CV applications”?
Yes, you can have a CNN that has a real continuous output.

What I meant applications other than computer-vision. E.g. say if you have structured data consisting of multiple features as separate columns. would each feature then be effectively a separate channel with only 1 value ( equivalent to 1 pixel in image recognition example)?

Yes, individual features can be used.