Bidirectional Long-Range Parser for Sequential Data Understanding

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Leotescu, George, Voinea, Daniel, Popa, Alin-Ionut
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914744166776832
author Leotescu, George
Voinea, Daniel
Popa, Alin-Ionut
author_facet Leotescu, George
Voinea, Daniel
Popa, Alin-Ionut
contents The transformer is a powerful data modelling framework responsible for remarkable performance on a wide range of tasks. However, they are limited in terms of scalability as it is suboptimal and inefficient to process long-sequence data. To this purpose we introduce BLRP (Bidirectional Long-Range Parser), a novel and versatile attention mechanism designed to increase performance and efficiency on long-sequence tasks. It leverages short and long range heuristics in the form of a local sliding window approach combined with a global bidirectional latent space synthesis technique. We show the benefits and versatility of our approach on vision and language domains by demonstrating competitive results against state-of-the-art methods on the Long-Range-Arena and CIFAR benchmarks together with ablations demonstrating the computational efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2404_05210
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Bidirectional Long-Range Parser for Sequential Data Understanding
Leotescu, George
Voinea, Daniel
Popa, Alin-Ionut
Computer Vision and Pattern Recognition
Computation and Language
Machine Learning
The transformer is a powerful data modelling framework responsible for remarkable performance on a wide range of tasks. However, they are limited in terms of scalability as it is suboptimal and inefficient to process long-sequence data. To this purpose we introduce BLRP (Bidirectional Long-Range Parser), a novel and versatile attention mechanism designed to increase performance and efficiency on long-sequence tasks. It leverages short and long range heuristics in the form of a local sliding window approach combined with a global bidirectional latent space synthesis technique. We show the benefits and versatility of our approach on vision and language domains by demonstrating competitive results against state-of-the-art methods on the Long-Range-Arena and CIFAR benchmarks together with ablations demonstrating the computational efficiency.
title Bidirectional Long-Range Parser for Sequential Data Understanding
topic Computer Vision and Pattern Recognition
Computation and Language
Machine Learning
url https://arxiv.org/abs/2404.05210