will deformable parts? will graphical models? will SVMs? There are trends that occur in ML and vision that greatly increase state-of-the-art at the time, deep learning is no different. I'm not saying they'll die and we'll all forget, but there's plenty of examples of new ideas succeeding over the favorite baseline.
Comments
previous discussion: https://news.ycombinator.com/item?id=9951352
will deformable parts? will graphical models? will SVMs? There are trends that occur in ML and vision that greatly increase state-of-the-art at the time, deep learning is no different. I'm not saying they'll die and we'll all forget, but there's plenty of examples of new ideas succeeding over the favorite baseline.
No. Not soon.