Implicit Reasoning in Deep Time Series Forecasting
This addresses the need to evaluate reasoning in time series models for researchers and practitioners, but it is incremental as it takes an initial step in an underexplored area.
The paper tackled the problem of assessing whether deep time series forecasting models truly understand temporal dynamics or just memorize data, finding that certain models generalize effectively in out-of-distribution scenarios, suggesting reasoning capabilities beyond memorization.
Recently, time series foundation models have shown promising zero-shot forecasting performance on time series from a wide range of domains. However, it remains unclear whether their success stems from a true understanding of temporal dynamics or simply from memorizing the training data. While implicit reasoning in language models has been studied, similar evaluations for time series models have been largely unexplored. This work takes an initial step toward assessing the reasoning abilities of deep time series forecasting models. We find that certain linear, MLP-based, and patch-based Transformer models generalize effectively in systematically orchestrated out-of-distribution scenarios, suggesting underexplored reasoning capabilities beyond simple pattern memorization.