Aurora: Towards Universal Generative Multimodal Time Series Forecasting

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wu, Xingjian, Jin, Jianxin, Qiu, Wanghui, Chen, Peng, Shu, Yang, Yang, Bin, Guo, Chenjuan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911497531162624
author Wu, Xingjian
Jin, Jianxin
Qiu, Wanghui
Chen, Peng
Shu, Yang
Yang, Bin
Guo, Chenjuan
author_facet Wu, Xingjian
Jin, Jianxin
Qiu, Wanghui
Chen, Peng
Shu, Yang
Yang, Bin
Guo, Chenjuan
contents Cross-domain generalization is very important in Time Series Forecasting because similar historical information may lead to distinct future trends due to the domain-specific characteristics. Recent works focus on building unimodal time series foundation models and end-to-end multimodal supervised models. Since domain-specific knowledge is often contained in modalities like texts, the former lacks the explicit utilization of them, thus hindering the performance. The latter is tailored for end-to-end scenarios and does not support zero-shot inference for cross-domain scenarios. In this work, we introduce Aurora, a Multimodal Time Series Foundation Model, which supports multimodal inputs and zero-shot inference. Pretrained on Cross-domain Multimodal Time Series Corpus, Aurora can adaptively extract and focus on key domain knowledge contained in corresponding text or image modalities, thus possessing strong cross-domain generalization capability. Through tokenization, encoding, and distillation, Aurora can extract multimodal domain knowledge as guidance and then utilizes a Modality-Guided Multi-head Self-Attention to inject them into the modeling of temporal representations. In the decoding phase, the multimodal representations are used to generate the conditions and prototypes of future tokens, contributing to a novel Prototype-Guided Flow Matching for generative probabilistic forecasting. Comprehensive experiments on 5 well-recognized benchmarks, including TimeMMD, TSFM-Bench, ProbTS, TFB, and EPF, demonstrate the consistent state-of-the-art performance of Aurora on both unimodal and multimodal scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2509_22295
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Aurora: Towards Universal Generative Multimodal Time Series Forecasting
Wu, Xingjian
Jin, Jianxin
Qiu, Wanghui
Chen, Peng
Shu, Yang
Yang, Bin
Guo, Chenjuan
Machine Learning
Cross-domain generalization is very important in Time Series Forecasting because similar historical information may lead to distinct future trends due to the domain-specific characteristics. Recent works focus on building unimodal time series foundation models and end-to-end multimodal supervised models. Since domain-specific knowledge is often contained in modalities like texts, the former lacks the explicit utilization of them, thus hindering the performance. The latter is tailored for end-to-end scenarios and does not support zero-shot inference for cross-domain scenarios. In this work, we introduce Aurora, a Multimodal Time Series Foundation Model, which supports multimodal inputs and zero-shot inference. Pretrained on Cross-domain Multimodal Time Series Corpus, Aurora can adaptively extract and focus on key domain knowledge contained in corresponding text or image modalities, thus possessing strong cross-domain generalization capability. Through tokenization, encoding, and distillation, Aurora can extract multimodal domain knowledge as guidance and then utilizes a Modality-Guided Multi-head Self-Attention to inject them into the modeling of temporal representations. In the decoding phase, the multimodal representations are used to generate the conditions and prototypes of future tokens, contributing to a novel Prototype-Guided Flow Matching for generative probabilistic forecasting. Comprehensive experiments on 5 well-recognized benchmarks, including TimeMMD, TSFM-Bench, ProbTS, TFB, and EPF, demonstrate the consistent state-of-the-art performance of Aurora on both unimodal and multimodal scenarios.
title Aurora: Towards Universal Generative Multimodal Time Series Forecasting
topic Machine Learning
url https://arxiv.org/abs/2509.22295