Can Transformers Learn n-gram Language Models?
Year: 2024
Type: conference-paper
Abstract: Much theoretical work has described the ability of transformers to represent formal languages. However, linking theoretical results to empirical performance is not straightforward due to the complex interplay between the architecture, the learning algorithm, and training data. To test whether theoretical lower bounds imply learnability of formal languages, we turn to recent work relating transformers to n-gram language models (LMs). We study transformers' ability to learn random n-gram LMs of two kinds: ones with arbitrary next-symbol probabilities and ones where those are defined with shared parameters. We find that classic estimation techniques for n-gram LMs such as add-lambda smoothing outperform transformers on the former, while transformers perform better on the latter, outperforming methods specifically designed to learn n-gram LMs.
Authors: Svete, Anej, Borenstein, Nadav, Zhou, Mike, Augenstein, Isabelle, Cotterell, Ryan
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6891315387
FWCI:
Cites:
Cited by:
Related to: