The Algorithms

A Brief Summary

Kempelen Consensus Approach: Miami
Network Deep Learning Algorithms Abstract

A Summary Overview of Deep Learning and Neural Network Techniques to Influence Blockchain Security, Speed, and Control

Abstract Summary Overview

Miami Network v0.14

The Kempelen consensus engine (PoSA-K) is an upgrade over the existing Parlia consensus engine deployed by the BNB chain. While keeping the lightweight, common-trust based staked authority proof validating consensus, it addresses the issues that may arise from the human decided authority assignment. For example, a shutdown of a chain utilizing Parlia can happen if the authority assigner team, and the existing validator nodes are removed. The Kempelen consensus engine proposes a distributed, intelligent authority assignment process that eliminates most of the human involvement in the upkeep of the chain.

GOALS

Deploying a consensus algorithm that assigns authority to a varying number of validators without any key kept on existing signers.

Implementing an automated statistical ZK key checker that can consistently provide an NP-complete query to prospecting validators in a fully deterministic environment (blockchain).

Implementing a routing protocol that fetches the node that suits the user’s transaction patterns, allowing to batch the transaction types on the same node, for optimized processing on the validator machines.

Summary of Features

Routing between nodes
for optimization:

Each node has a running score
of transaction processing time
by transaction type.
Transactions are batched by type onto each block.
These batches are pushed into the highest scored nodes.

Join the Miami Revolution

New Developers

Ignite your next dApp on Miami.

[Learn more here]

Unite with Miami

Forge a partnership to
amplify our tech together?

[Reach more here]

Project Teams

Transition to Miami with
a few clicks.

[Learn more here]

Miami Foundation

$25m in grants and sponsorships for projects
bolstering Web3 and the Miami ecosystem.

[Reach more here]

Stay in the Eye
of the Storm

Share your email, and let's team up

Summary of Features

Routing between nodes for optimization:

Each node has a running score of transaction processing time
by transaction type

Transactions are batched by type onto each block.

These batches are pushed into the highest scored nodes.


Hamiltonian Statistical Zero Knowledge
Automated Authority Assignment

The blockchain state is kept as a continuous hamiltonian graph.

Node 1 is an existing validator, Node 2 is a prospector.

A set of transactions in P are assigned (by a learning model) as nodes on a
hamiltonian cycle ‘G’.

Node 1 generates a matrix H isomorphic to G

Depending on which authority pool the prospector is trying to join two separate
queries are formed. Node 2 will either have to prove that it knows the isomorphism
between H and G by demonstration, or it will need to provide the number of vertices
of H.

Node 1 has this hamiltonian cycle (received by the AI key generation node, with
varying difficulty depending on the blockchain’s needs).

The computing time for both of these operations through the means of brute force
is too high, however, if Node 2 can provide a passing answer to either question, it will
be granted signing authority over a fixed amount of blocks, then this process is
repeated.



Therefore removing the need for a password to be kept on a centralized database,
and suited for a decentralized network.


Routing protocol is observed and upgraded by a distributed
intelligent agent that can track the transactor’s (sender wallet)
tx pattern, and allows a faster indexing to which transaction
block the incoming tx belongs.