A HyperNetwork


Our Mission is to build a scalable production system HyperNetwork for federated learning across carrier networks. HyperNetwork.

Problem: Networks are operated independently, despite having similar infrastructure and following standardized protocols. This is expensive and redundant, as problems and operations tasks repeat from one network to another.

We believe that it is possible to maintain each networks privacy while also building an AI that can learn from each network to prevent faults and optimize all networks collectively and more intelligently – creating a “HyperNetwork”.

Solution: A HyperNetwork is built on Decentralized Network In-situ Data. Local AI Agents Push updates, like weights, to the Global AI “HyperNetwork” where their Collective Intelligence is then Pushed Back to Each Node.

The solution allows carriers to collaboratively learn AI models while keeping all their network data private.

Blank Diagram - Page 2 (1).png