Zero-shot Multivariate Time Series Forecasting Using Tabular Prior Fitted Networks
Mayuka Jayawardhana, Nihal Sharma, Kazem Meidani, Bayan Bruss, Tom Goldstein, & Doron Bergman · 2026
Read Zero-shot Multivariate Time Series Forecasting Using Tabular Prior Fitted Networks by Mayuka Jayawardhana, Nihal Sharma, Kazem Meidani, Bayan Bruss, Tom Goldstein, & Doron Bergman (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-08400.txt.
Description
Jayawardhana and five colleagues turn multivariate time series forecasting into a tabular problem that foundation models can solve without task specific training. Earlier tabular approaches split the channels apart and forecast each alone, throwing away links between channels. Their framework instead recasts the whole task as a chain of single value regressions, solved zero shot by any tabular model with regression skill, and tests with TabPFN-TS match leading tabular rivals.
Catalog record
| CREATOR | Mayuka Jayawardhana, Nihal Sharma, Kazem Meidani, Bayan Bruss, Tom Goldstein, & Doron Bergman |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21272 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.08400