Intrinsic Lorentz Neural Network
Xianglong Shi, Ziheng Chen, Yunhan Jiang, Nicu Sebe · 2026
Read Intrinsic Lorentz Neural Network by Xianglong Shi, Ziheng Chen, Yunhan Jiang, Nicu Sebe (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-2602-23981.txt.
Description
Xianglong Shi and three co-authors present the Intrinsic Lorentz Neural Network, a fully intrinsic hyperbolic architecture for data with hidden hierarchical structure. The 2026 paper notes that earlier hyperbolic networks mix Euclidean steps with hyperbolic ones or rely on extrinsic parameterizations, breaking the geometry they claim to use.
At the core sits a point-to-hyperplane fully connected layer, which replaces Euclidean affine scores with closed-form hyperbolic distances from features to learned Lorentz hyperplanes, so decision boundaries respect curvature. Around it the authors build intrinsic modules. One module is a Lorentz batch normalization that couples centering with scaling. Another is a gyro-additive bias. A third is a patch-concatenation operator that aligns expected log-radius across feature blocks. A fourth is a Lorentz dropout layer. Tests on image and genomic benchmarks show strong gains over prior hyperbolic designs. Every step stays inside the curved space claimed.
Catalog record
| CREATOR | Xianglong Shi, Ziheng Chen, Yunhan Jiang, Nicu Sebe |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 76 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-23981