Mercurial > pylearn
comparison sparse_random_autoassociator/main.py @ 371:22463a194c90
Update doc
author | Joseph Turian <turian@gmail.com> |
---|---|
date | Mon, 07 Jul 2008 01:57:49 -0400 |
parents | a1bbcde6b456 |
children | 75bab24bb2d8 |
comparison
equal
deleted
inserted
replaced
370:a1bbcde6b456 | 371:22463a194c90 |
---|---|
5 | 5 |
6 The learned model is:: | 6 The learned model is:: |
7 h = sigmoid(dot(x, w1) + b1) | 7 h = sigmoid(dot(x, w1) + b1) |
8 y = sigmoid(dot(h, w2) + b2) | 8 y = sigmoid(dot(h, w2) + b2) |
9 | 9 |
10 We assume that most of the inputs are zero, and hence that we can | 10 We assume that most of the inputs are zero, and hence that |
11 separate x into xnonzero, x's nonzero components, and a xzero, | 11 we can separate x into xnonzero, x's nonzero components, and |
12 a sample of the zeros. (We randomly without replacement choose | 12 xzero, a sample of the zeros. We sample---randomly without |
13 ZERO_SAMPLE_SIZE zero columns.) | 13 replacement---ZERO_SAMPLE_SIZE zero columns from x. |
14 | 14 |
15 The desideratum is that every nonzero entry is separated from every | 15 The desideratum is that every nonzero entry is separated from every |
16 zero entry by margin at least MARGIN. | 16 zero entry by margin at least MARGIN. |
17 For each ynonzero, we want it to exceed max(yzero) by at least MARGIN. | 17 For each ynonzero, we want it to exceed max(yzero) by at least MARGIN. |
18 For each yzero, we want it to be exceed by min(ynonzero) by at least MARGIN. | 18 For each yzero, we want it to be exceed by min(ynonzero) by at least MARGIN. |