Communication Network Clustering Using VAE-Stratified SOMs

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing clustering algorithms in communication networks are dependent on parameters and hyperparameters, leading to inconsistent results and increased operational costs due to iterative convergence, and lack domain agnosticism and efficiency in data clustering.

Innovation Solution

Implementing automated machine learning models with domain knowledge for record recognition and linkage, using vector embeddings and self-organizing maps with stratified sampling to autonomously cluster network operational data, enabling efficient and domain-agnostic data clustering.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If existing clustering algorithms are used with parameters and hyperparameters, then clustering can be performed, but the results are inconsistent and operational costs increase due to iterative convergence

Engineering Contradiction:

Improveclustering result consistencyVSAvoidoperational efficiency

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system uses automated machine learning to enable the clustering algorithm to automatically select and optimize its own parameters without human intervention. The self-service mechanism eliminates manual parameter tuning and iterative convergence processes, resulting in both consistent results and reduced operational costs.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent transforms the clustering approach by changing from traditional parameter-dependent algorithms to automated machine learning models that dynamically adjust parameters. This parameter transformation enables the system to adapt to different data types and domains while maintaining consistent clustering results without manual intervention.

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If traditional clustering algorithms are used, then clustering can be performed, but they lack domain agnosticism and efficiency

Engineering Contradiction:

Improvedomain agnosticismVSAvoiddata clustering efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent implements universal automated machine learning models that can perform clustering across multiple domains and data types without requiring domain-specific customization. The multi-functional system processes diverse network operational data, customer service data, and other heterogeneous data types using the same clustering framework, achieving both domain agnosticism and efficiency.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20260005956A1Automatic clustering-based communication network management

Publication Date: 2026.01.01 AT&T INTELLECTUAL PROPERTY I L P

AI summary

A processing system may generate vector embeddings from network operational data of a communication network. The processing system may next apply a variational autoencoder to the vector embeddings to create a set of stratified samples, where the set of stratified samples comprises at least a first portion of the plurality of vector embeddings. In addition, the processing system may train a self-organizing map using the stratified samples to create a plurality of clusters. The processing system may next apply the self-organizing map to at least a second portion of the plurality of vector embeddings to assign the at least the second portion to respective clusters of the plurality of clusters. The processing system may then identify at least one characteristic associated with at least one cluster and may perform at least one remedial action in the communication network in response to the identifying of the at least one characteristic.