TiVaT: A Transformer with a Single Unified Mechanism for Capturing Asynchronous Dependencies in Multivariate Time Series Forecasting

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ha, Junwoo, Kwon, Hyukjae, Kim, Sungsoo, Lee, Kisu, Park, Seungjae, Kim, Ha Young
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910806555230208
author Ha, Junwoo
Kwon, Hyukjae
Kim, Sungsoo
Lee, Kisu
Park, Seungjae
Kim, Ha Young
author_facet Ha, Junwoo
Kwon, Hyukjae
Kim, Sungsoo
Lee, Kisu
Park, Seungjae
Kim, Ha Young
contents Multivariate time series (MTS) forecasting is vital across various domains but remains challenging due to the need to simultaneously model temporal and inter-variate dependencies. Existing channel-dependent models, where Transformer-based models dominate, process these dependencies separately, limiting their capacity to capture complex interactions such as lead-lag dynamics. To address this issue, we propose TiVaT (Time-variate Transformer), a novel architecture incorporating a single unified module, a Joint-Axis (JA) attention module, that concurrently processes temporal and variate modeling. The JA attention module dynamically selects relevant features to particularly capture asynchronous interactions. In addition, we introduce distance-aware time-variate sampling in the JA attention, a novel mechanism that extracts significant patterns through a learned 2D embedding space while reducing noise. Extensive experiments demonstrate TiVaT's overall performance across diverse datasets, particularly excelling in scenarios with intricate asynchronous dependencies.
format Preprint
id arxiv_https___arxiv_org_abs_2410_01531
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle TiVaT: A Transformer with a Single Unified Mechanism for Capturing Asynchronous Dependencies in Multivariate Time Series Forecasting
Ha, Junwoo
Kwon, Hyukjae
Kim, Sungsoo
Lee, Kisu
Park, Seungjae
Kim, Ha Young
Machine Learning
Artificial Intelligence
I.2.0
Multivariate time series (MTS) forecasting is vital across various domains but remains challenging due to the need to simultaneously model temporal and inter-variate dependencies. Existing channel-dependent models, where Transformer-based models dominate, process these dependencies separately, limiting their capacity to capture complex interactions such as lead-lag dynamics. To address this issue, we propose TiVaT (Time-variate Transformer), a novel architecture incorporating a single unified module, a Joint-Axis (JA) attention module, that concurrently processes temporal and variate modeling. The JA attention module dynamically selects relevant features to particularly capture asynchronous interactions. In addition, we introduce distance-aware time-variate sampling in the JA attention, a novel mechanism that extracts significant patterns through a learned 2D embedding space while reducing noise. Extensive experiments demonstrate TiVaT's overall performance across diverse datasets, particularly excelling in scenarios with intricate asynchronous dependencies.
title TiVaT: A Transformer with a Single Unified Mechanism for Capturing Asynchronous Dependencies in Multivariate Time Series Forecasting
topic Machine Learning
Artificial Intelligence
I.2.0
url https://arxiv.org/abs/2410.01531