Unified Training of Universal Time Series Forecasting Transformers

TL;DR

This paper introduces Moirai, a universal time series forecasting Transformer trained on a large dataset, capable of handling diverse tasks without dataset-specific tuning, outperforming traditional models.

Contribution

The paper presents Moirai, a novel Transformer architecture for universal time series forecasting, trained on LOTSA, addressing cross-frequency, multivariate, and distributional challenges.

Findings

Moirai achieves competitive zero-shot forecasting performance.

Trained on LOTSA with over 27 billion observations.

Outperforms traditional models on diverse datasets.

Abstract

Deep learning for time series forecasting has traditionally operated within a one-model-per-dataset framework, limiting its potential to leverage the game-changing impact of large pre-trained models. The concept of universal forecasting, emerging from pre-training on a vast collection of time series datasets, envisions a single Large Time Series Model capable of addressing diverse downstream forecasting tasks. However, constructing such a model poses unique challenges specific to time series data: i) cross-frequency learning, ii) accommodating an arbitrary number of variates for multivariate time series, and iii) addressing the varying distributional properties inherent in large-scale data. To address these challenges, we present novel enhancements to the conventional time series Transformer architecture, resulting in our proposed Masked Encoder-based Universal Time Series Forecasting Transformer (Moirai). Trained on our newly introduced Large-scale Open Time Series Archive (LOTSA) featuring over 27B observations across nine domains, Moirai achieves competitive or superior performance as a zero-shot forecaster when compared to full-shot models. Code, data, and model weights can be found at https://github.com/SalesforceAIResearch/uni2ts.