From 1b8181bf0d6e9137e6b9ccdbe414aec37377a1a9 Mon Sep 17 00:00:00 2001
From: Chris Dyer <cdyer@allegro.clab.cs.cmu.edu>
Date: Sun, 18 Nov 2012 13:35:42 -0500
Subject: major restructure of the training code

---
 training/utils/risk.cc | 45 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 45 insertions(+)
 create mode 100644 training/utils/risk.cc

(limited to 'training/utils/risk.cc')

diff --git a/training/utils/risk.cc b/training/utils/risk.cc
new file mode 100644
index 00000000..d5a12cfd
--- /dev/null
+++ b/training/utils/risk.cc
@@ -0,0 +1,45 @@
+#include "risk.h"
+
+#include "prob.h"
+#include "candidate_set.h"
+#include "ns.h"
+
+using namespace std;
+
+namespace training {
+
+// g = \sum_e p(e|f) * loss(e) * (phi(e,f) - E[phi(e,f)])
+double CandidateSetRisk::operator()(const vector<double>& params,
+                                    SparseVector<double>* g) const {
+  prob_t z;
+  for (unsigned i = 0; i < cands_.size(); ++i) {
+    const prob_t u(cands_[i].fmap.dot(params), init_lnx());
+    z += u;
+  }
+  const double log_z = log(z);
+
+  SparseVector<double> exp_feats;
+  if (g) {
+    for (unsigned i = 0; i < cands_.size(); ++i) {
+      const double log_prob = cands_[i].fmap.dot(params) - log_z;
+      const double prob = exp(log_prob);
+      exp_feats += cands_[i].fmap * prob;
+    }
+  }
+
+  double risk = 0;
+  for (unsigned i = 0; i < cands_.size(); ++i) {
+    const double log_prob = cands_[i].fmap.dot(params) - log_z;
+    const double prob = exp(log_prob);
+    const double cost = metric_.IsErrorMetric() ? metric_.ComputeScore(cands_[i].eval_feats)
+                                                : 1.0 - metric_.ComputeScore(cands_[i].eval_feats);
+    const double r = prob * cost;
+    risk += r;
+    if (g) (*g) += (cands_[i].fmap - exp_feats) * r;
+  }
+  return risk;
+}
+
+}
+
+
-- 
cgit v1.2.3