Saved in:
Bibliographic Details
Main Authors: Hayakawa, Daichi, Sato, Issei
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.12413
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929546374152192
author Hayakawa, Daichi
Sato, Issei
author_facet Hayakawa, Daichi
Sato, Issei
contents In this study, we provide constructive proof that Transformers can recognize and generate hierarchical language efficiently with respect to model size, even without the need for a specific positional encoding. Specifically, we show that causal masking and a starting token enable Transformers to compute positional information and depth within hierarchical structures. We demonstrate that Transformers without positional encoding can generate hierarchical languages. Furthermore, we suggest that explicit positional encoding might have a detrimental effect on generalization with respect to sequence length.
format Preprint
id arxiv_https___arxiv_org_abs_2410_12413
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Theoretical Analysis of Hierarchical Language Recognition and Generation by Transformers without Positional Encoding
Hayakawa, Daichi
Sato, Issei
Computation and Language
In this study, we provide constructive proof that Transformers can recognize and generate hierarchical language efficiently with respect to model size, even without the need for a specific positional encoding. Specifically, we show that causal masking and a starting token enable Transformers to compute positional information and depth within hierarchical structures. We demonstrate that Transformers without positional encoding can generate hierarchical languages. Furthermore, we suggest that explicit positional encoding might have a detrimental effect on generalization with respect to sequence length.
title Theoretical Analysis of Hierarchical Language Recognition and Generation by Transformers without Positional Encoding
topic Computation and Language
url https://arxiv.org/abs/2410.12413