23 F eb 2 01 6 Learning Efficient Algorithms with Hierarchical Attentive Memory