The reviewed record of science sign in
Pith

arxiv: 2506.03128 · v1 · pith:EKZBWSEH · submitted 2025-06-03 · cs.LG

Zero-Shot Time Series Forecasting with Covariates via In-Context Learning

Andreas Auer , Raghul Parthipan , Pedro Mercado , Abdul Fatir Ansari , Lorenzo Stella , Bernie Wang , Michael Bohlke-Schneider , Syama Sundar Rangapuram This is my paper

Reviewed by Pithpith:EKZBWSEHopen to challenge →

classification cs.LG
keywords covariatesforecastingzero-shotcosmicseriestimeeffectivelyin-context
0
0 comments X
read the original abstract

Pretrained time series models, capable of zero-shot forecasting, have demonstrated significant potential in enhancing both the performance and accessibility of time series forecasting. However, existing pretrained models either do not support covariates or fail to incorporate them effectively. We introduce COSMIC, a zero-shot forecasting model that utilizes covariates via in-context learning. To address the challenge of data scarcity, we propose Informative Covariate Augmentation, which enables the training of COSMIC without requiring any datasets that include covariates. COSMIC achieves state-of-the-art performance in zero-shot forecasting, both with and without covariates. Our quantitative and qualitative analysis demonstrates that COSMIC effectively leverages covariates in zero-shot forecasting.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Why Do Time Series Models Need Long Context Windows?

    cs.LG 2026-06 unverdicted novelty 7.0

    Long input windows are required to identify the generative process in time series forecasting even for short-memory processes, and decoupling identification from forecasting improves scalability.

  2. RMISC: A Large-scale Real-world Multivariate Corpus for Time Series Foundation Models

    cs.AI 2026-07 conditional novelty 6.0

    A curated 142-billion-point real-world multivariate time series corpus improves zero-shot forecasting when combined with existing synthetic and univariate pretraining data across four foundation models.

  3. TiRex-2: Generalizing TiRex to Multivariate Data and Streaming

    cs.LG 2026-07 unverdicted novelty 6.0

    TiRex-2 is a recurrent xLSTM time series foundation model for multivariate forecasting with future covariates and constant-cost streaming that reports SOTA zero-shot results on GIFT-Eval and fev-bench.

  4. CITRAS-FM: Tiny Time Series Foundation Model for Covariate-Informed Zero-Shot Forecasting

    cs.LG 2026-06 unverdicted novelty 4.0

    CITRAS-FM is a 7M-param decoder-only Transformer TSFM with Shifted Attention and CovSynth synthetic covariate pretraining that claims SOTA zero-shot accuracy among sub-10M models on fev-bench with sub-0.1s CPU inference.