TVNet: A novel time series analysis method based on dynamic convolution and 3D-Variation
Chenghan Li, Mingchen Li & Ruisheng Diao · 2025
Read TVNet: A novel time series analysis method based on dynamic convolution and 3D-Variation by Chenghan Li, Mingchen Li & Ruisheng Diao (2025) 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-2503-07674.txt.
Description
A convolution net rebuilt for time series is set out by Chenghan Li, Mingchen Li, and Ruisheng Diao in a 2025 arXiv paper. Transform and mix nets have led the field of late, while plain convolution nets lagged. Their TVNet reshapes series into a 3D form that keeps patch, in-patch, and cross-channel links. Dynamic filters then read that form with light cost.
Across five core tasks, the net topped the charts of its day. It held the speed edge that made convolution nets dear. Moves to new sets and rough data both held up well. Old net styles gain fresh life here.
Catalog record
| CREATOR | Chenghan Li, Mingchen Li & Ruisheng Diao |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 96 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-07674