Position: The Inevitable End of One-Architecture-Fits-All-Domains in Time Series Forecasting
Qinwei Ma, Jingzhe Shi, Jiahao Qiu, Zaiwen Yang · 2026
Read Position: The Inevitable End of One-Architecture-Fits-All-Domains in Time Series Forecasting by Qinwei Ma, Jingzhe Shi, Jiahao Qiu, Zaiwen Yang (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-2602-01736.txt.
Description
Qinwei Ma and colleagues argue that one architecture cannot fit all domains in time series forecasting, as the field hits its limits. General domain architectures grow ever more complex while performance saturates. They offer little to real practice in specific domains such as finance, weather and traffic, where each domain builds its own methods. The authors note that domain methods rarely use recent neural network advances from the time series community. They call for a shift: either deep learning for specific domains, or meta learning for general domains.
Catalog record
| CREATOR | Qinwei Ma, Jingzhe Shi, Jiahao Qiu, Zaiwen Yang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37968 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.01736