TempoPFN: Synthetic Pre-training of Linear RNNs for Zero-shot Time Series Forecasting [R]

Accepted at NeurIPS 2026 TLDR: The paper presents a library of synthetic time series generators and augmentations to train time series foundation models, in their case a DeltaProduct based linear RNN. The synthetic data has already been used in Toto-2 and TS-ICL.