PENGUIN: Enhancing Transformer with Periodic-Nested Group Attention for Long-term Time Series Forecasting

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sun, Tian, Chen, Yuqi, Sun, Weiwei
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908918207217664
author Sun, Tian
Chen, Yuqi
Sun, Weiwei
author_facet Sun, Tian
Chen, Yuqi
Sun, Weiwei
contents Despite advances in the Transformer architecture, their effectiveness for long-term time series forecasting (LTSF) remains controversial. In this paper, we investigate the potential of integrating explicit periodicity modeling into the self-attention mechanism to enhance the performance of Transformer-based architectures for LTSF. Specifically, we propose PENGUIN, a simple yet effective periodic-nested group attention mechanism. Our approach introduces a periodic-aware relative attention bias to directly capture periodic structures and a grouped multi-query attention mechanism to handle multiple coexisting periodicities (e.g., daily and weekly cycles) within time series data. Extensive experiments across diverse benchmarks demonstrate that PENGUIN consistently outperforms both MLP-based and Transformer-based models. Code is available at https://github.com/ysygMhdxw/AISTATS2026_PENGUIN.
format Preprint
id arxiv_https___arxiv_org_abs_2508_13773
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PENGUIN: Enhancing Transformer with Periodic-Nested Group Attention for Long-term Time Series Forecasting
Sun, Tian
Chen, Yuqi
Sun, Weiwei
Machine Learning
Artificial Intelligence
Despite advances in the Transformer architecture, their effectiveness for long-term time series forecasting (LTSF) remains controversial. In this paper, we investigate the potential of integrating explicit periodicity modeling into the self-attention mechanism to enhance the performance of Transformer-based architectures for LTSF. Specifically, we propose PENGUIN, a simple yet effective periodic-nested group attention mechanism. Our approach introduces a periodic-aware relative attention bias to directly capture periodic structures and a grouped multi-query attention mechanism to handle multiple coexisting periodicities (e.g., daily and weekly cycles) within time series data. Extensive experiments across diverse benchmarks demonstrate that PENGUIN consistently outperforms both MLP-based and Transformer-based models. Code is available at https://github.com/ysygMhdxw/AISTATS2026_PENGUIN.
title PENGUIN: Enhancing Transformer with Periodic-Nested Group Attention for Long-term Time Series Forecasting
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2508.13773