Saved in:
Bibliographic Details
Main Authors: Wang, Xue, Zhou, Tian, Gao, Jinyang, Ding, Bolin, Zhou, Jingren
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.11029
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909647299936256
author Wang, Xue
Zhou, Tian
Gao, Jinyang
Ding, Bolin
Zhou, Jingren
author_facet Wang, Xue
Zhou, Tian
Gao, Jinyang
Ding, Bolin
Zhou, Jingren
contents We present a joint forecasting framework for time series prediction that contrasts with traditional direct or recursive methods. This framework achieves state-of-the-art performance for our designed foundation model, YingLong, and reveals a novel scaling effect: longer outputs significantly enhance model accuracy due to delayed chain-of-thought reasoning in our non-causal approach. YingLong is a non-causal, bidirectional attention encoder-only transformer trained through masked token recovery, aligning more effectively with language understanding tasks than with generation tasks. Additionally, we boost performance by tackling output variance with a multi-input ensemble. We release four foundation models ranging from 6M to 300M parameters, demonstrating superior results in zero-shot tasks on the ETT and Weather datasets. YingLong achieves more than 60% best performance. To ensure generalizability, we assessed the models using the GIFT-Eval benchmark, which comprises 23 time series datasets across 7 domains. Yinglong significantly outperformed the best time-series foundation models, end-to-end trained models by 14% and 44% in rank respectively.The pretrained 300M model is available at https://huggingface.co/qcw1314/YingLong_300m
format Preprint
id arxiv_https___arxiv_org_abs_2506_11029
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Output Scaling: YingLong-Delayed Chain of Thought in a Large Pretrained Time Series Forecasting Model
Wang, Xue
Zhou, Tian
Gao, Jinyang
Ding, Bolin
Zhou, Jingren
Machine Learning
Artificial Intelligence
We present a joint forecasting framework for time series prediction that contrasts with traditional direct or recursive methods. This framework achieves state-of-the-art performance for our designed foundation model, YingLong, and reveals a novel scaling effect: longer outputs significantly enhance model accuracy due to delayed chain-of-thought reasoning in our non-causal approach. YingLong is a non-causal, bidirectional attention encoder-only transformer trained through masked token recovery, aligning more effectively with language understanding tasks than with generation tasks. Additionally, we boost performance by tackling output variance with a multi-input ensemble. We release four foundation models ranging from 6M to 300M parameters, demonstrating superior results in zero-shot tasks on the ETT and Weather datasets. YingLong achieves more than 60% best performance. To ensure generalizability, we assessed the models using the GIFT-Eval benchmark, which comprises 23 time series datasets across 7 domains. Yinglong significantly outperformed the best time-series foundation models, end-to-end trained models by 14% and 44% in rank respectively.The pretrained 300M model is available at https://huggingface.co/qcw1314/YingLong_300m
title Output Scaling: YingLong-Delayed Chain of Thought in a Large Pretrained Time Series Forecasting Model
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2506.11029