Are Language Models Actually Useful for Time Series Forecasting?
This work challenges the assumption that LLMs are beneficial for time series forecasting, potentially saving computational resources for researchers and practitioners in this domain.
The paper investigates the utility of large language models (LLMs) in time series forecasting, finding that removing or replacing LLMs with simpler attention layers does not degrade and often improves performance, with no benefits in computational cost, representation, or few-shot settings.
Large language models (LLMs) are being applied to time series forecasting. But are language models actually useful for time series? In a series of ablation studies on three recent and popular LLM-based time series forecasting methods, we find that removing the LLM component or replacing it with a basic attention layer does not degrade forecasting performance -- in most cases, the results even improve! We also find that despite their significant computational cost, pretrained LLMs do no better than models trained from scratch, do not represent the sequential dependencies in time series, and do not assist in few-shot settings. Additionally, we explore time series encoders and find that patching and attention structures perform similarly to LLM-based forecasters.