Context-Free Transductions with Neural Stacks

Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP

Venue: BlackboxNLP
Type: Workshop
Interpretability
Formal Languages
Architectures
Authors
Affiliation

Sophie Hao

Yale University

William Merrill

Yale University

Dana Angluin

Yale University

Robert Frank

Yale University

Noah Amsel

Yale University

Andrew Benz

Yale University

Simon Mendelsohn

Yale University

Published

November 1, 2018

Abstract
This paper analyzes the behavior of stack-augmented recurrent neural network (RNN) models. Due to the architectural similarity between stack RNNs and pushdown transducers, we train stack RNN models on a number of tasks, including string reversal, context-free language modelling, and cumulative XOR evaluation. Examining the behavior of our networks, we show that stack-augmented RNNs can discover intuitive stack-based strategies for solving our tasks. However, stack RNNs are more difficult to train than classical architectures such as LSTMs. Rather than employ stack-based strategies, more complex networks often find approximate solutions by using the stack as unstructured memory.