Retrieval Augmented Time Series Forecasting

TL;DR

This paper introduces Retrieval Augmented Forecasting (RAF), a framework that enhances time-series forecasting by retrieving related data, demonstrating improved accuracy especially with larger models.

Contribution

It adapts retrieval-augmented generation techniques to time-series forecasting, providing a new framework and strategies that improve forecast accuracy across domains.

Findings

RAF improves forecasting accuracy across diverse domains.

Benefits of RAF are more significant for larger models.

Efficient retrieval strategies are key to RAF's success.

Abstract

Retrieval-augmented generation (RAG) is a central component of modern LLM systems, particularly in scenarios where up-to-date information is crucial for accurately responding to user queries or when queries exceed the scope of the training data. The advent of time-series foundation models (TSFM), such as Chronos, and the need for effective zero-shot forecasting performance across various time-series domains motivates the question: Do benefits of RAG similarly carry over to time series forecasting? In this paper, we advocate that the dynamic and event-driven nature of time-series data makes RAG a crucial component of TSFMs and introduce a principled RAG framework for time-series forecasting, called Retrieval Augmented Forecasting (RAF). Within RAF, we develop efficient strategies for retrieving related time-series examples and incorporating them into forecast. Through experiments and mechanistic studies, we demonstrate that RAF indeed improves the forecasting accuracy across diverse time series domains and the improvement is more significant for larger TSFM sizes.