diff --git a/MetaAugment/autoaugment_learners/aa_learner.py b/MetaAugment/autoaugment_learners/aa_learner.py
index a562a3dcedab4dc7e6b86a921e87ef53a03f572b..00d38aab9b292b1a3e88aae9d76310fbc3fdc370 100644
--- a/MetaAugment/autoaugment_learners/aa_learner.py
+++ b/MetaAugment/autoaugment_learners/aa_learner.py
@@ -394,15 +394,16 @@ class aa_learner:
             first_trans, first_prob, first_mag = subpol[0]
             second_trans, second_prob, second_mag = subpol[1]
             components = (first_prob, first_mag, second_prob, second_mag)
-            if second_trans in pol_dict[first_trans]:
-                pol_dict[first_trans][second_trans].append(components)
+            if first_trans in pol_dict:
+                if second_trans in pol_dict[first_trans]:
+                    pol_dict[first_trans][second_trans].append(components)
+                else:
+                    pol_dict[first_trans]= {second_trans: [components]}
             else:
                 pol_dict[first_trans]= {second_trans: [components]}
         self.policy_record[curr_pol] = (pol_dict, accuracy)
 
         self.num_pols_tested += 1
-        
-
         return accuracy
     
 
diff --git a/MetaAugment/autoaugment_learners/rand_augment_learner.py b/MetaAugment/autoaugment_learners/rand_augment_learner.py
new file mode 100644
index 0000000000000000000000000000000000000000..b6974befe7438f91db36c37cc1a6b5c15801813f
--- /dev/null
+++ b/MetaAugment/autoaugment_learners/rand_augment_learner.py
@@ -0,0 +1,8 @@
+import torch 
+import numpy as np
+from MetaAugment.autoaugment_learners.randomsearch_learner import randomsearch_learner
+
+class rand_augment_learner(randomsearch_learner):
+
+    def __init__(self):
+        pass
\ No newline at end of file