The Neural State Pushdown Automata

TL;DR

This paper introduces the Neural State Pushdown Automaton (NSPDA), a neural network with a digital stack that effectively recognizes context-free grammars, improves training convergence with prior rule knowledge, and advances understanding of neural automata.

Contribution

The paper presents the NSPDA model with a digital stack, a noise regularization scheme for tensor networks, and a method for incorporating grammar rules to enhance training and generalization.

Findings

NSPDA effectively recognizes context-free grammars.

Prior rule knowledge significantly improves convergence and generalization.

Compared to other RNNs, NSPDA shows superior performance on CFGs.

Abstract

In order to learn complex grammars, recurrent neural networks (RNNs) require sufficient computational resources to ensure correct grammar recognition. A widely-used approach to expand model capacity would be to couple an RNN to an external memory stack. Here, we introduce a "neural state" pushdown automaton (NSPDA), which consists of a digital stack, instead of an analog one, that is coupled to a neural network state machine. We empirically show its effectiveness in recognizing various context-free grammars (CFGs). First, we develop the underlying mechanics of the proposed higher order recurrent network and its manipulation of a stack as well as how to stably program its underlying pushdown automaton (PDA) to achieve desired finite-state network dynamics. Next, we introduce a noise regularization scheme for higher-order (tensor) networks, to our knowledge the first of its kind, and design an algorithm for improved incremental learning. Finally, we design a method for inserting grammar rules into a NSPDA and empirically show that this prior knowledge improves its training convergence time by an order of magnitude and, in some cases, leads to better generalization. The NSPDA is also compared to a classical analog stack neural network pushdown automaton (NNPDA) as well as a wide array of first and second-order RNNs with and without external memory, trained using different learning algorithms. Our results show that, for Dyck(2) languages, prior rule-based knowledge is critical for optimization convergence and for ensuring generalization to longer sequences at test time. We observe that many RNNs with and without memory, but no prior knowledge, fail to converge and generalize poorly on CFGs.