Lorentzian Logic: Solving Unknown Cluster Numbers via Differentiable Graph Entropy

Written by hyperbole | Published 2026/02/14
Tech Story Tags: deep-learning | graph-clustering | hyperbolic-space | lsenet-neural-net | lorentz-convolution | graph-conductance | lorentz | lorentzian-logic

TLDR

Optimize network analysis with LSEnet, a deep learning model that eliminates the need to predefined cluster numbers. By leveraging the Lorentz model of hyperbolic space and differentiable structural information, it identifies optimal hierarchies through self-organized gradient backpropagation.via the TL;DR App

Table of Links

Abstract and 1. Introduction

B. Hyperbolic Space

C. Technical Details

D. Additional Results

5.1. Embedding Leaf Nodes

where ⊙ is the Hadamard product, masking the attentional weight if the corresponding edge does not exist in the graph.

Authors:

(1) Li Sun, North China Electric Power University, Beijing 102206, China ([email protected]);

(2) Zhenhao Huang, North China Electric Power University, Beijing 102206, China;

(3) Hao Peng, Beihang University, Beijing 100191, China;

(4) Yujie Wang, North China Electric Power University, Beijing 102206, China;

(5) Chunyang Liu, Didi Chuxing, Beijing, China;

(6) Philip S. Yu, University of Illinois at Chicago, IL, USA.

This paper is available on arxiv under CC BY-NC-SA 4.0 Deed (Attribution-Noncommercial-Sharelike 4.0 International) license.

Written by hyperbole | Amplifying words and ideas to separate the ordinary from the extraordinary, making the mundane majestic.

Published by HackerNoon on 2026/02/14