Beyond Self Attention: A Subquadratic Fourier Wavelet Transformer with Multi Modal Fusion

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kiruluta, Andrew, Lemos, Andreas, Lundy, Eric
Format: Preprint
Published: 2021
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915253509423104
author Kiruluta, Andrew
Lemos, Andreas
Lundy, Eric
author_facet Kiruluta, Andrew
Lemos, Andreas
Lundy, Eric
contents We revisit the use of spectral techniques to replaces the attention mechanism in Transformers through Fourier Transform based token mixing, and present a comprehensive and novel reformulation of this technique in next generation transformer models. We provide expanded literature context, detailed mathematical formulations of Fourier mixing and causal masking, and introduce a novel MultiDomain Fourier Wavelet Attention(MDFWA) that integrates frequency and time localized transforms to capture both global and local dependencies efficiently. We derive the complexity bounds, gradient formulas, and show that MDFWA achieves sub quadratic time and memory cost while improving expressive power. We validate our design on an abstractive summarization task using PubMed dataset, by enhancing the proposed approach with learned frequency bases, adaptive scale selection, and multi-modal extensions.
format Preprint
id arxiv_https___arxiv_org_abs_2111_15473
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Beyond Self Attention: A Subquadratic Fourier Wavelet Transformer with Multi Modal Fusion
Kiruluta, Andrew
Lemos, Andreas
Lundy, Eric
Computation and Language
Machine Learning
We revisit the use of spectral techniques to replaces the attention mechanism in Transformers through Fourier Transform based token mixing, and present a comprehensive and novel reformulation of this technique in next generation transformer models. We provide expanded literature context, detailed mathematical formulations of Fourier mixing and causal masking, and introduce a novel MultiDomain Fourier Wavelet Attention(MDFWA) that integrates frequency and time localized transforms to capture both global and local dependencies efficiently. We derive the complexity bounds, gradient formulas, and show that MDFWA achieves sub quadratic time and memory cost while improving expressive power. We validate our design on an abstractive summarization task using PubMed dataset, by enhancing the proposed approach with learned frequency bases, adaptive scale selection, and multi-modal extensions.
title Beyond Self Attention: A Subquadratic Fourier Wavelet Transformer with Multi Modal Fusion
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2111.15473