StretchTime: Adaptive Time Series Forecasting via Symplectic Attention
Yubin Kim, Viresh Pati, Jevon Twitty, Vinh Pham, Shihao Yang, Jiecheng Lu · 2026
Read StretchTime: Adaptive Time Series Forecasting via Symplectic Attention by Yubin Kim, Viresh Pati, Jevon Twitty, Vinh Pham, Shihao Yang, Jiecheng Lu (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-2602-08983.txt.
Description
Clocks in the real world stretch and squeeze: financial cycles shift, biological rhythms flex, yet transformers assume time ticks evenly. Yubin Kim and five co-authors first prove that rotary position embeddings cannot represent non-affine time warping at all. Their answer is Symplectic Positional Embeddings, a learnable encoding drawn from Hamiltonian mechanics that generalizes rotations to the symplectic group.
An input-dependent warp module lets attention dilate or contract time coordinates end to end, capturing locally varying periodicities without preset warp functions. Built into StretchTime, a multivariate forecaster, the method reaches state-of-the-art scores on standard benchmarks and proves sturdier on non-stationary data.
Catalog record
| CREATOR | Yubin Kim, Viresh Pati, Jevon Twitty, Vinh Pham, Shihao Yang, Jiecheng Lu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 63 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-08983