Learning Bounded Context-Free-Grammar via LSTM and the Transformer:Difference and Explanations

TL;DR

This paper compares LSTM and Transformer models in learning context-free grammars, revealing that Transformers better capture stack operations due to their latent space decomposition, explaining their superior performance.

Contribution

It introduces an oracle training paradigm to analyze how LSTM and Transformer decompose latent spaces and their ability to simulate stack operations in CFL learning.

Findings

Transformers outperform LSTMs in representing stack operations without forced decomposition.

Forced decomposition aligns LSTM and Transformer performance in CFL learning.

Transformers' latent space better captures stack-based automaton transitions.

Abstract

Long Short-Term Memory (LSTM) and Transformers are two popular neural architectures used for natural language processing tasks. Theoretical results show that both are Turing-complete and can represent any context-free language (CFL).In practice, it is often observed that Transformer models have better representation power than LSTM. But the reason is barely understood. We study such practical differences between LSTM and Transformer and propose an explanation based on their latent space decomposition patterns. To achieve this goal, we introduce an oracle training paradigm, which forces the decomposition of the latent representation of LSTM and the Transformer and supervises with the transitions of the Pushdown Automaton (PDA) of the corresponding CFL. With the forced decomposition, we show that the performance upper bounds of LSTM and Transformer in learning CFL are close: both of them can simulate a stack and perform stack operation along with state transitions. However, the absence of forced decomposition leads to the failure of LSTM models to capture the stack and stack operations, while having a marginal impact on the Transformer model. Lastly, we connect the experiment on the prototypical PDA to a real-world parsing task to re-verify the conclusions