Dana, the author's book dates back several years, which is a concern since the entire field has undergone an explosion in new work since maybe 2012 or so--at least at some point after Hinton's team won the ImageNet competition, and after Bengio's approach to NLP started turning heads. Both of those advancements were based on neural nets. Have you heard whether researchers have used the kernel memory approach successfully, perhaps to beat a state-of-the-art benchmark in a research paper or competition?
You are viewing a single comment's thread from: