Enhancing Time Series Classification with Diversity-Driven Neural Network Ensembles
Javidan Abdullayev, Maxime Devanne, Cyril Meyer, Ali Ismail-Fawaz, Jonathan Weber, Germain Forestier · 2026
Read Enhancing Time Series Classification with Diversity-Driven Neural Network Ensembles by Javidan Abdullayev, Maxime Devanne, Cyril Meyer, Ali Ismail-Fawaz, Jonathan Weber, Germain Forestier (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-2602-07579.txt.
Description
Ensembles of neural networks for time series classification usually train identical models and hope diversity emerges, but identical training breeds redundant traits. Javidan Abdullayev and five co-authors build a diversity-driven ensemble that explicitly encourages members to learn other traits.
A decorrelated learning strategy applies a feature orthogonality loss to learned representations, pushing ensemble members apart in feature space so their combination adds real information. The method targets time series classification, where ensembles already lead but often waste their members. Users get an ensemble method that earns its extra models through enforced diversity rather than mere replication.
Catalog record
| CREATOR | Javidan Abdullayev, Maxime Devanne, Cyril Meyer, Ali Ismail-Fawaz, Jonathan Weber, Germain Forestier |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07579