Work

Method and Apparatus for Relativistic Manifold Embedding and Normalization in Deep Neural Networks

Year: 2026

Type: article

Abstract: This technical report specifies the mathematical formulation and architectural design for Relativistic Manifold Embedding and Normalization in deep neural networks. It introduces Minkowski RMSNorm, dynamic hyperboloid coordinate projection mechanics, analytical timelike coordinate derivation, and a hybrid exact/asymptotic logarithmic geodesic distance matrix calculation algorithm on a Pseudo-Riemannian Lorentz Hyperboloid manifold. ---Author Links & Resources:- Personal Website: https://baar.uk/- Hugging Face Profile: https://huggingface.co/aixk

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.21775576

Author: Oh, Youngjin

Language: English

Topic: Neural Networks and Applications

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7172338328

FWCI:

Cites:

Cited by:

Related to: