What is the "hashing trick"? - MetaOptimize Q+A
Let's say we want to design a function v = phi(x), which from a d-dimensional vector x = (x(1), x(2), ..., x(d)) outputs a new m-dimensional vector v, with m either greater or smaller than d. In other words, phi can be used either for reducing dimensionality of x (d > m) or for sparsifying x (m > d). One way to do so is to use a hash function h to map x(1) to v(h(1)), x(2) to v(h(2)), ..., x(d) to
Hashing Representations
I've discovered that there are many 'hashing tricks' in machine learning. Some of these are like the count-min sketch in that they rely on an explicit bloom filter style datastructure. The ones here take a radical conceptual step: they only use one hashing function. This turns out to work out remarkably well when learning, because the learning algorithm can learn to deal with collisions. In severa
Text Classification and Feature Hashing: Sparse Matrix-Vector Multiplication with Cython | Newsle Blog
by Erich Owens Optimizing the memory footprint of a classifier used here at Newsle set us down a rabbit hole of rewriting a basic Scipy function with Cython, something that only became a problem when our high-dimensional text spaces grew to a cartoonish size, thanks to the hashing trick. Here I motivate the use of the hashing trick, how we use sparse matrix-vector multiplication for text classific
1
リリース、障害情報などのサービスのお知らせ
最新の人気エントリーの配信
処理を実行中です
j次のブックマーク
k前のブックマーク
lあとで読む
eコメント一覧を開く
oページを開く
PFI Seminar 2012/03/15 カーネルとハッシュの機械学習
