-
Notifications
You must be signed in to change notification settings - Fork 1.9k
/
autolink.cc
46 lines (37 loc) · 1.23 KB
/
autolink.cc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
#include "reductions.h"
const int autoconstant = 524267083;
struct autolink
{ uint32_t d; // degree of the polynomial
uint32_t stride_shift;
};
template <bool is_learn>
void predict_or_learn(autolink& b, LEARNER::base_learner& base, example& ec)
{ base.predict(ec);
float base_pred = ec.pred.scalar;
// add features of label
ec.indices.push_back(autolink_namespace);
features& fs = ec.feature_space[autolink_namespace];
for (size_t i = 0; i < b.d; i++)
if (base_pred != 0.)
{ fs.push_back(base_pred, autoconstant + (i << b.stride_shift));
base_pred *= ec.pred.scalar;
}
ec.total_sum_feat_sq += fs.sum_feat_sq;
if (is_learn)
base.learn(ec);
else
base.predict(ec);
ec.total_sum_feat_sq -= fs.sum_feat_sq;
fs.erase();
ec.indices.pop();
}
LEARNER::base_learner* autolink_setup(vw& all)
{ if (missing_option<size_t, true>(all, "autolink", "create link function with polynomial d"))
return nullptr;
autolink& data = calloc_or_throw<autolink>();
data.d = (uint32_t)all.vm["autolink"].as<size_t>();
data.stride_shift = all.weights.stride_shift();
LEARNER::learner<autolink>& ret =
init_learner(&data, setup_base(all), predict_or_learn<true>, predict_or_learn<false>);
return make_base(ret);
}