dtrain: simplified pair generation

2 files changed, 16 insertions, 140 deletions

diff --git a/training/dtrain/dtrain.cc b/training/dtrain/dtrain.cc
index 18addcb0..69630206 100644
--- a/training/dtrain/dtrain.cc
+++ b/training/dtrain/dtrain.cc
@@ -21,9 +21,7 @@ dtrain_init(int argc, char** argv, po::variables_map* conf)
     ("epochs",            po::value<unsigned>()->default_value(10),                               "# of iterations T (per shard)")
     ("k",                 po::value<unsigned>()->default_value(100),                            "how many translations to sample")
     ("filter",            po::value<string>()->default_value("uniq"),                          "filter kbest list: 'not', 'uniq'")
-    ("pair_sampling",     po::value<string>()->default_value("XYX"),                 "how to sample pairs: 'all', 'XYX' or 'PRO'")
     ("hi_lo",             po::value<float>()->default_value(0.1),                   "hi and lo (X) for XYX (default 0.1), <= 0.5")
-    ("pair_threshold",    po::value<score_t>()->default_value(0.),                         "bleu [0,1] threshold to filter pairs")
     ("N",                 po::value<unsigned>()->default_value(4),                                          "N for Ngrams (BLEU)")
     ("scorer",            po::value<string>()->default_value("stupid_bleu"),      "scoring: bleu, stupid_, smooth_, approx_, lc_")
     ("learning_rate",     po::value<weight_t>()->default_value(1.0),                                              "learning rate")
@@ -34,7 +32,6 @@ dtrain_init(int argc, char** argv, po::variables_map* conf)
     ("l1_reg_strength",   po::value<weight_t>(),                                                     "l1 regularization strength")
     ("fselect",           po::value<weight_t>()->default_value(-1), "select top x percent (or by threshold) of features after each epoch NOT IMPLEMENTED") // TODO
     ("loss_margin",       po::value<weight_t>()->default_value(0.),  "update if no error in pref pair but model scores this near")
-    ("max_pairs",         po::value<unsigned>()->default_value(std::numeric_limits<unsigned>::max()), "max. # of pairs per Sent.")
     ("pclr",              po::value<string>()->default_value("no"),         "use a (simple|adagrad) per-coordinate learning rate")
     ("batch",             po::value<bool>()->zero_tokens(),                                               "do batch optimization")
     ("repeat",            po::value<unsigned>()->default_value(1),     "repeat optimization over kbest list this number of times")
@@ -56,14 +53,6 @@ dtrain_init(int argc, char** argv, po::variables_map* conf)
     cerr << cl << endl;
     return false;
   }
-  if ((*conf)["pair_sampling"].as<string>() != "all" && (*conf)["pair_sampling"].as<string>() != "XYX" &&
-        (*conf)["pair_sampling"].as<string>() != "PRO" && (*conf)["pair_sampling"].as<string>() != "output_pairs") {
-    cerr << "Wrong 'pair_sampling' param: '" << (*conf)["pair_sampling"].as<string>() << "'." << endl;
-    return false;
-  }
-  if (conf->count("hi_lo") && (*conf)["pair_sampling"].as<string>() != "XYX") {
-    cerr << "Warning: hi_lo only works with pair_sampling XYX." << endl;
-  }
   if ((*conf)["hi_lo"].as<float>() > 0.5 || (*conf)["hi_lo"].as<float>() < 0.01) {
     cerr << "hi_lo must lie in [0.01, 0.5]" << endl;
     return false;
@@ -72,10 +61,6 @@ dtrain_init(int argc, char** argv, po::variables_map* conf)
     cerr << "No training data given." << endl;
     return false;
   }
-  if ((*conf)["pair_threshold"].as<score_t>() < 0) {
-    cerr << "The threshold must be >= 0!" << endl;
-    return false;
-  }
   if ((*conf)["select_weights"].as<string>() != "last" && (*conf)["select_weights"].as<string>() != "best" &&
         (*conf)["select_weights"].as<string>() != "avg" && (*conf)["select_weights"].as<string>() != "VOID") {
     cerr << "Wrong 'select_weights' param: '" << (*conf)["select_weights"].as<string>() << "', use 'last' or 'best'." << endl;
@@ -106,12 +91,9 @@ main(int argc, char** argv)
   const unsigned N = conf["N"].as<unsigned>();
   const unsigned T = conf["epochs"].as<unsigned>();
   const unsigned stop_after = conf["stop_after"].as<unsigned>();
-  const string pair_sampling = conf["pair_sampling"].as<string>();
-  const score_t pair_threshold = conf["pair_threshold"].as<score_t>();
   const string select_weights = conf["select_weights"].as<string>();
   const string output_ranking = conf["output_ranking"].as<string>();
   const float hi_lo = conf["hi_lo"].as<float>();
-  const unsigned max_pairs = conf["max_pairs"].as<unsigned>();
   int repeat = conf["repeat"].as<unsigned>();
   weight_t loss_margin = conf["loss_margin"].as<weight_t>();
   bool batch = false;
@@ -192,17 +174,13 @@ main(int argc, char** argv)
     cerr << setw(25) << "gamma " << gamma << endl;
     cerr << setw(25) << "loss margin " << loss_margin << endl;
     cerr << setw(25) << "faster perceptron " << faster_perceptron << endl;
-    cerr << setw(25) << "pairs " << "'" << pair_sampling << "'" << endl;
-    if (pair_sampling == "XYX")
-      cerr << setw(25) << "hi lo " << hi_lo << endl;
-    cerr << setw(25) << "pair threshold " << pair_threshold << endl;
+    cerr << setw(25) << "hi lo " << hi_lo << endl;
     cerr << setw(25) << "select weights " << "'" << select_weights << "'" << endl;
     if (conf.count("l1_reg"))
       cerr << setw(25) << "l1 reg " << l1_reg << " '" << conf["l1_reg"].as<string>() << "'" << endl;
     if (rescale)
       cerr << setw(25) << "rescale " << rescale << endl;
     cerr << setw(25) << "pclr " << pclr << endl;
-    cerr << setw(25) << "max pairs " << max_pairs << endl;
     cerr << setw(25) << "repeat " << repeat << endl;
     cerr << setw(25) << "cdec conf " << "'" << conf["decoder_config"].as<string>() << "'" << endl;
     cerr << setw(25) << "input " << "'" << input_fn << "'" << endl;
@@ -335,28 +313,12 @@ main(int argc, char** argv)
     if (!noup) {
       // get pairs
       vector<pair<ScoredHyp,ScoredHyp> > pairs;
-      if (pair_sampling == "all")
-        all_pairs(samples, pairs, pair_threshold, max_pairs, faster_perceptron);
-      if (pair_sampling == "XYX")
-        partXYX(samples, pairs, pair_threshold, max_pairs, faster_perceptron, hi_lo);
-      if (pair_sampling == "PRO")
-        PROsampling(samples, pairs, pair_threshold, max_pairs);
-      if (pair_sampling == "output_pairs")
-        all_pairs(samples, pairs, pair_threshold, max_pairs, false);
+      MakePairs(samples, pairs, faster_perceptron, hi_lo);
       int cur_npairs = pairs.size();
       npairs += cur_npairs;
 
       score_t kbest_loss_first = 0.0, kbest_loss_last = 0.0;
 
-      if (pair_sampling == "output_pairs") {
-        for (auto p: pairs) {
-          cout << p.first.model << " ||| "  << p.first.score << " ||| " <<  p.first.f  << endl;
-          cout << p.second.model << " ||| "  << p.second.score << " ||| " <<  p.second.f  << endl;
-          cout << endl;
-        }
-        continue;
-      }
-
       for (vector<pair<ScoredHyp,ScoredHyp> >::iterator it = pairs.begin();
            it != pairs.end(); it++) {
         if (rescale) {
diff --git a/training/dtrain/pairs.h b/training/dtrain/pairs.h
index fd08be8c..dea0dabc 100644
--- a/training/dtrain/pairs.h
+++ b/training/dtrain/pairs.h
@@ -1,140 +1,54 @@
-#ifndef _DTRAIN_PAIRSAMPLING_H_
-#define _DTRAIN_PAIRSAMPLING_H_
+#ifndef _DTRAIN_PAIRS_H_
+#define _DTRAIN_PAIRS_H_
 
 namespace dtrain
 {
 
-
 bool
-accept_pair(score_t a, score_t b, score_t threshold)
-{
-  if (fabs(a - b) < threshold) return false;
-  return true;
-}
-
-bool
-cmp_hyp_by_score_d(ScoredHyp a, ScoredHyp b)
+CmpHypsByScore(ScoredHyp a, ScoredHyp b)
 {
   return a.score > b.score;
 }
 
-inline void
-all_pairs(vector<ScoredHyp>* s, vector<pair<ScoredHyp,ScoredHyp> >& training, score_t threshold, unsigned max, bool misranked_only, float _unused=1)
-{
-  sort(s->begin(), s->end(), cmp_hyp_by_score_d);
-  unsigned sz = s->size();
-  bool b = false;
-  unsigned count = 0;
-  for (unsigned i = 0; i < sz-1; i++) {
-    for (unsigned j = i+1; j < sz; j++) {
-      if (misranked_only && !((*s)[i].model <= (*s)[j].model)) continue;
-      if (threshold > 0) {
-        if (accept_pair((*s)[i].score, (*s)[j].score, threshold))
-          training.push_back(make_pair((*s)[i], (*s)[j]));
-      } else {
-        if ((*s)[i].score != (*s)[j].score)
-          training.push_back(make_pair((*s)[i], (*s)[j]));
-      }
-      if (++count == max) {
-        b = true;
-        break;
-      }
-    }
-    if (b) break;
-  }
-}
-
 /*
  * multipartite ranking
  *  sort (descending) by bleu
- *  compare top X to middle Y and low X
+ *  compare top X (hi) to middle Y (med) and low X (lo)
  *  cmp middle Y to low X
  */
-
 inline void
-partXYX(vector<ScoredHyp>* s, vector<pair<ScoredHyp,ScoredHyp> >& training, score_t threshold, unsigned max, bool misranked_only, float hi_lo)
+MakePairs(vector<ScoredHyp>* s,
+          vector<pair<ScoredHyp,ScoredHyp> >& training,
+          bool misranked_only,
+          float hi_lo)
 {
   unsigned sz = s->size();
   if (sz < 2) return;
-  sort(s->begin(), s->end(), cmp_hyp_by_score_d);
+  sort(s->begin(), s->end(), CmpHypsByScore);
   unsigned sep = round(sz*hi_lo);
+  // hi vs. med vs. low
   unsigned sep_hi = sep;
   if (sz > 4) while (sep_hi < sz && (*s)[sep_hi-1].score == (*s)[sep_hi].score) ++sep_hi;
   else sep_hi = 1;
-  bool b = false;
-  unsigned count = 0;
   for (unsigned i = 0; i < sep_hi; i++) {
     for (unsigned j = sep_hi; j < sz; j++) {
       if (misranked_only && !((*s)[i].model <= (*s)[j].model)) continue;
-      if (threshold > 0) {
-        if (accept_pair((*s)[i].score, (*s)[j].score, threshold))
-          training.push_back(make_pair((*s)[i], (*s)[j]));
-      } else {
-        if ((*s)[i].score != (*s)[j].score)
-          training.push_back(make_pair((*s)[i], (*s)[j]));
-      }
-      if (++count == max) {
-        b = true;
-        break;
-      }
+      if ((*s)[i].score != (*s)[j].score)
+        training.push_back(make_pair((*s)[i], (*s)[j]));
     }
-    if (b) break;
   }
+  // med vs. low
   unsigned sep_lo = sz-sep;
   while (sep_lo > 0 && (*s)[sep_lo-1].score == (*s)[sep_lo].score) --sep_lo;
   for (unsigned i = sep_hi; i < sep_lo; i++) {
     for (unsigned j = sep_lo; j < sz; j++) {
       if (misranked_only && !((*s)[i].model <= (*s)[j].model)) continue;
-      if (threshold > 0) {
-        if (accept_pair((*s)[i].score, (*s)[j].score, threshold))
-          training.push_back(make_pair((*s)[i], (*s)[j]));
-      } else {
-        if ((*s)[i].score != (*s)[j].score)
-          training.push_back(make_pair((*s)[i], (*s)[j]));
-      }
-      if (++count == max) return;
-    }
-  }
-}
-
-/*
- * pair sampling as in
- * 'Tuning as Ranking' (Hopkins & May, 2011)
- *     count = max (5000)
- * threshold = 5% BLEU (0.05 for param 3)
- *       cut = top 10%
- */
-bool
-_PRO_cmp_pair_by_diff_d(pair<ScoredHyp,ScoredHyp> a, pair<ScoredHyp,ScoredHyp> b)
-{
-  return (fabs(a.first.score - a.second.score)) > (fabs(b.first.score - b.second.score));
-}
-inline void
-PROsampling(vector<ScoredHyp>* s, vector<pair<ScoredHyp,ScoredHyp> >& training, score_t threshold, unsigned max, bool _unused=false, float _also_unused=0)
-{
-  sort(s->begin(), s->end(), cmp_hyp_by_score_d);
-  unsigned max_count = max, count = 0, sz = s->size();
-  bool b = false;
-  for (unsigned i = 0; i < sz-1; i++) {
-    for (unsigned j = i+1; j < sz; j++) {
-      if (accept_pair((*s)[i].score, (*s)[j].score, threshold)) {
+      if ((*s)[i].score != (*s)[j].score)
         training.push_back(make_pair((*s)[i], (*s)[j]));
-        if (++count == max_count) {
-          b = true;
-          break;
-        }
-      }
     }
-    if (b) break;
-  }
-  if (training.size() > max/10) {
-    sort(training.begin(), training.end(), _PRO_cmp_pair_by_diff_d);
-    training.erase(training.begin()+(max/10), training.end());
   }
-  return;
 }
 
-
 } // namespace
 
 #endif