MOMENT: A Family of Open Time-series Foundation Models

TL;DR

MOMENT introduces a family of open-source foundation models for time series analysis, leveraging a large, diverse dataset and a new benchmark to evaluate performance in limited supervision scenarios.

Contribution

The paper presents MOMENT, a novel approach to pre-training large-scale time series models using the comprehensive Time Series Pile and a new evaluation benchmark.

Findings

Models perform well with minimal data and fine-tuning.

Pre-training improves performance across diverse tasks.

Empirical observations reveal insights about large time series models.

Abstract

We introduce MOMENT, a family of open-source foundation models for general-purpose time series analysis. Pre-training large models on time series data is challenging due to (1) the absence of a large and cohesive public time series repository, and (2) diverse time series characteristics which make multi-dataset training onerous. Additionally, (3) experimental benchmarks to evaluate these models, especially in scenarios with limited resources, time, and supervision, are still in their nascent stages. To address these challenges, we compile a large and diverse collection of public time series, called the Time series Pile, and systematically tackle time series-specific challenges to unlock large-scale multi-dataset pre-training. Finally, we build on recent work to design a benchmark to evaluate time series foundation models on diverse tasks and datasets in limited supervision settings. Experiments on this benchmark demonstrate the effectiveness of our pre-trained models with minimal data and task-specific fine-tuning. Finally, we present several interesting empirical observations about large pre-trained time series models. Pre-trained models (AutonLab/MOMENT-1-large) and Time Series Pile (AutonLab/Timeseries-PILE) are available on Huggingface.