540 learning configurations. Need to pull gamma option in

bin/medley

@@ -38,6 +38,14 @@ def main():
         choices=["full", "single", "greedy"] 
     )
 
+    global_parser.add_argument(
+        "--reward",
+        help="the kind of reward to give",
+        type=str,
+        default="bump",
+        choices=["binary", "bump", "exp"] 
+    )
+
     global_parser.add_argument(
         "--classifier",
         "-c",
@@ -211,7 +219,7 @@ def main():
         with open(args.load_classifier, "rb") as f:
             classifier = dill.load(f)
 
-    execute(problems, args.output, classifier, timeout_manager, args.timeout, args.feature_setting, args.extra)
+    execute(problems, args.output, classifier, timeout_manager, args.timeout, args.feature_setting, args.extra, args.reward)
 
     if args.save_classifier:
         classifier.save(args.save_classifier)

clean.sh

@@ -0,0 +1,3 @@
+for learner in thompson neighbor nearest knearest random MLP linear exp3; do
+    rm $1/${learner}*
+done 

medleysolver/runner.py

@@ -6,7 +6,7 @@
 from medleysolver.distributions import ExponentialDist
 from medleysolver.dispatch import run_problem
 
-def execute(problems, output, classifier, time_manager, timeout, feature_setting, extra_time_to_first):
+def execute(problems, output, classifier, time_manager, timeout, feature_setting, extra_time_to_first, reward):
     mean = 0
     writer = csv.writer(open(output, 'w'))
 
@@ -29,15 +29,16 @@ def execute(problems, output, classifier, time_manager, timeout, feature_setting
         writer.writerow(solved_prob)
 
 
-def apply_ordering(problem, order, timeout, time_manager, extra_time_to_first, times):
+def apply_ordering(problem, order, timeout, time_manager, extra_time_to_first, times, reward):
     elapsed = 0
     rewards = [-1 for _ in SOLVERS] # negative rewards should be ignored. 
     time_spent = []
+    timeout = int(timeout)
 
     budgets = [int(time_manager.get_timeout(solver, times))+1 for solver in order]
 
     for i in range(len(budgets)):
-        budgets[i] = min(budgets[i], int(timeout - sum(budgets[:i])))
+        budgets[i] = min(budgets[i], max(0, int(timeout - sum(budgets[:i]))))
 
     if sum(budgets) < timeout:
         if extra_time_to_first:
@@ -54,7 +55,13 @@ def apply_ordering(problem, order, timeout, time_manager, extra_time_to_first, t
         time_for_solver = int(timeout - elapsed) + 1 if i == len(order) - 1 else budgets[i]
         res = run_problem(solver, SOLVERS[solver], problem, time_for_solver)
 
-        reward = 1 + ((1 - res.elapsed / timeout) ** 4) if is_solved(res.result) else 0
+        if reward == "binary":
+            reward = 1 if is_solved(res.result) else 0
+        elif reward == "bump":
+            reward = 1 + ((1 - res.elapsed / timeout) ** 4) if is_solved(res.result) else 0
+        elif reward == "exp":
+            reward = (1 - res.elapsed / timeout) ** 4 if is_solved(res.result) else 0
+
         rewards[list(SOLVERS.keys()).index(solver)] = reward
         time_spent.append(res.elapsed)
 

ordermisses.py

@@ -24,7 +24,7 @@
         for row in spamreader:
             r = float(row[3]) if isinstance(row[3], str) else row[3]
             r = r if r < 60 else 60
-            r = r if row[4] != "error" else 60
+            r = r if row[4] != "error" and "timeout" not in row[4] else 60
             cols[solver][row[0]] = r
 
 print("choice, answer, problem")

process.py

@@ -27,6 +27,6 @@
 totaltimes = [str(sum(time)) for time in times]
 totalcount = [str(sum(map(lambda x: 1 if x == "sat" or x == "unsat" else 0, answer))) for answer in answers]
 
-print(",".join(solvers))
-print(",".join(totaltimes))
-print(",".join(totalcount))
+print("solver, total time, total solved")
+for i in range(len(solvers)):
+    print(",".join([str(solvers[i]), str(totaltimes[i]), str(totalcount[i])]))

runmedley.sh

@@ -1,17 +1,26 @@
-medley ./$1/ ./$1/thompson.csv -c thompson
-medley ./$1/ ./$1/nearest.csv -c neighbor
-medley ./$1/ ./$1/knearest.csv -c knearest --k 10
-medley ./$1/ ./$1/random.csv -c random 
-medley ./$1/ ./$1/MLP.csv -c MLP 
-medley ./$1/ ./$1/linear.csv -c linear 
-medley ./$1/ ./$1/exp3.csv -c exp3 
-
-medley ./$1/ ./$1/thompson_only.csv -c thompson --timeout_manager const --set_const 60
-medley ./$1/ ./$1/nearest_only.csv -c neighbor --timeout_manager const --set_const 60
-medley ./$1/ ./$1/knearest_only.csv -c knearest --k 10 --timeout_manager const --set_const 60
-medley ./$1/ ./$1/random_only.csv -c random --timeout_manager const --set_const 60
-medley ./$1/ ./$1/MLP_only.csv -c MLP --timeout_manager const --set_const 60
-medley ./$1/ ./$1/linear_only.csv -c linear --timeout_manager const --set_const 60
-medley ./$1/ ./$1/exp3_only.csv -c exp3 --timeout_manager const --set_const 60
-
-medley ./$1/ ./$1/timenearest.csv -c knearest --k 10 --timeout_manager nearest --time_k 20
+for seed in 0 1; do
+    for learner in thompson neighbor knearest random MLP linear exp3a exp3b exp3c exp3d; do
+        if [ $learner == exp3a ]
+        then
+            learnconfig="exp3 --gamma 0.07"
+        elif [ $learner == exp3b ]
+        then
+            learnconfig="exp3 --gamma 0.1"
+        elif [ $learner == exp3c ]
+        then
+            learnconfig="exp3 --gamma 0.25"
+        elif [ $learner == exp3c ]
+        then
+            learnconfig="exp3 --gamma 0.5"
+        else 
+            learnconfig=$learner
+        fi
+        for feature in both probes bow; do
+            for reward in binary bump exp; do
+                medley ./$1/ ./$1/${learner}_${feature}_${reward}_const_${seed}.csv   --classifier $learnconfig --seed $seed --feature_setting $feature --reward $reward --timeout_manager const --set_const 60
+                medley ./$1/ ./$1/${learner}_${feature}_${reward}_expo_${seed}.csv    --classifier $learnconfig --seed $seed --feature_setting $feature --reward $reward --timeout_manager expo
+                medley ./$1/ ./$1/${learner}_${feature}_${reward}_nearest_${seed}.csv --classifier $learnconfig --seed $seed --feature_setting $feature --reward $reward --timeout_manager nearest --time_k 20
+            done 
+        done 
+    done 
+done 

timemisses.py

@@ -24,7 +24,7 @@
         for row in spamreader:
             r = float(row[3]) if isinstance(row[3], str) else row[3]
             r = r if r < 60 else 60
-            r = r if row[4] != "error" else 60
+            r = r if row[4] != "error" and "timeout" not in row[4] else 60
             cols[solver][row[0]] = r
 
 print("solver, answer, guess, problem")