{"id":482,"date":"2008-12-07T19:46:22","date_gmt":"2008-12-08T01:46:22","guid":{"rendered":"http:\/\/hunch.net\/?p=482"},"modified":"2008-12-07T19:46:22","modified_gmt":"2008-12-08T01:46:22","slug":"a-nips-paper","status":"publish","type":"post","link":"https:\/\/hunch.net\/?p=482","title":{"rendered":"A NIPS paper"},"content":{"rendered":"

I’m skipping NIPS this year in favor of Ada<\/a>, but I wanted to point out this paper<\/a> by Andriy Mnih<\/a> and Geoff Hinton<\/a>. The basic claim of the paper is that by carefully but automatically constructing a binary tree over words, it’s possible to predict words well with huge computational resource savings over unstructured approaches.<\/p>\n

I’m interested in this beyond the application to word prediction because it is relevant to the general normalization problem: If you want to predict the probability of one of a large number of events, often you must compute a predicted score for all the events and then normalize, a computationally inefficient operation. The problem comes up in many places using probabilistic models, but I’ve run into it with high-dimensional regression.<\/p>\n

There are a couple workarounds for this computational bug:<\/p>\n

    \n
  1. Approximate. There are many ways. Often the approximations are uncontrolled (i.e. can be arbitrarily bad), and hence finicky in application.<\/li>\n
  2. Avoid. You don’t really want a probability, you want the most probable choice which can be found more directly. Energy based model<\/a> update rules are an example of that approach and there are many other direct methods from supervised learning. This is great when it applies, but sometimes a probability is actually needed.<\/li>\n<\/ol>\n

    This paper points out that a third approach can be viable empirically: use a self-normalizing structure. It seems highly likely that this is true in other applications as well.<\/p>\n","protected":false},"excerpt":{"rendered":"

    I’m skipping NIPS this year in favor of Ada, but I wanted to point out this paper by Andriy Mnih and Geoff Hinton. The basic claim of the paper is that by carefully but automatically constructing a binary tree over words, it’s possible to predict words well with huge computational resource savings over unstructured approaches. … <\/p>\n

    Continue reading “A NIPS paper”<\/span><\/a><\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6,27,29,18],"tags":[],"class_list":["post-482","post","type-post","status-publish","format-standard","hentry","category-bayesian","category-empirical","category-machine-learning","category-papers"],"_links":{"self":[{"href":"https:\/\/hunch.net\/index.php?rest_route=\/wp\/v2\/posts\/482","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/hunch.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/hunch.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/hunch.net\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/hunch.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=482"}],"version-history":[{"count":0,"href":"https:\/\/hunch.net\/index.php?rest_route=\/wp\/v2\/posts\/482\/revisions"}],"wp:attachment":[{"href":"https:\/\/hunch.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=482"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hunch.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=482"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hunch.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=482"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}