This is a relatively short overview of deep learning; the opening paragraphs help a lot with the definition of the term, and make clear its relationship to artificial intelligence in general. It doesn't go into a lot of depth (for that, you will want Schmidhuber (33 page PDF)). At the end of the paper there's an interesting discussion of the limitations of deep learning based on adversarial attacks. These are "indicative of imperfect representations adopted by DL systems. While humans are able to reason, infer, and extrapolate from observations, these capabilities have not been well developed in DL systems.
Today: 2 Total: 8 [Direct link] [Share]
