diff --git a/README.md b/README.md
index 8ac0d3e..295b257 100644
--- a/README.md
+++ b/README.md
@@ -29,6 +29,6 @@ PUBLICATIONS
 
 The segmentation methods are described in:
 
-Rico Sennrich, Barry Haddow and Alexandra Birch (2015):
+Rico Sennrich, Barry Haddow and Alexandra Birch (2016):
     Neural Machine Translation of Rare Words with Subword Units
-    http://arxiv.org/abs/1508.07909
\ No newline at end of file
+    Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL 2016). Berlin, Germany.
\ No newline at end of file
diff --git a/apply_bpe.py b/apply_bpe.py
index 736e822..bb0d7c8 100755
--- a/apply_bpe.py
+++ b/apply_bpe.py
@@ -7,7 +7,8 @@ The text will not be smaller, but use only a fixed vocabulary, with rare words
 encoded as variable-length sequences of subword units.
 
 Reference:
-Rico Sennrich, Barry Haddow and Alexandra Birch (2015). Neural Machine Translation of Rare Words with Subword Units.
+Rico Sennrich, Barry Haddow and Alexandra Birch (2016). Neural Machine Translation of Rare Words with Subword Units.
+Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL 2016). Berlin, Germany.
 """
 
 from __future__ import unicode_literals, division
diff --git a/bpe_toy.py b/bpe_toy.py
index 6e9bed3..967236f 100644
--- a/bpe_toy.py
+++ b/bpe_toy.py
@@ -9,7 +9,8 @@ This is an (inefficient) toy implementation that shows the algorithm. For proces
 indexing and incremental updates can be used to speed up the implementation (see learn_bpe.py).
 
 Reference:
-Rico Sennrich, Barry Haddow and Alexandra Birch (2015). Neural Machine Translation of Rare Words with Subword Units.
+Rico Sennrich, Barry Haddow and Alexandra Birch (2016). Neural Machine Translation of Rare Words with Subword Units.
+Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL 2016). Berlin, Germany.
 """
 
 
diff --git a/learn_bpe.py b/learn_bpe.py
index 22e8782..8c9f293 100755
--- a/learn_bpe.py
+++ b/learn_bpe.py
@@ -7,7 +7,8 @@ Unlike the original BPE, it does not compress the plain text, but can be used to
 of a text to a configurable number of symbols, with only a small increase in the number of tokens.
 
 Reference:
-Rico Sennrich, Barry Haddow and Alexandra Birch (2015). Neural Machine Translation of Rare Words with Subword Units.
+Rico Sennrich, Barry Haddow and Alexandra Birch (2016). Neural Machine Translation of Rare Words with Subword Units.
+Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL 2016). Berlin, Germany.
 """
 
 from __future__ import unicode_literals