Ai-enhanced entropy measurement system for randomness evaluation
AssigneeEntrokey Labs Inc.
InventorScott STREIT
Life of this patent
Estimating the term…
Abstract
A method and system for estimating the entropy of bit sequences using a hybrid deep learning model. The system employs a neural network model that integrates statistical and pattern-based analysis to assess entropy. A min-entropy estimation function determines randomness levels, while a convolutional neural network (CNN) extracts sequential patterns. The system leverages a structured training approach with configurable parameters, allowing for adaptive entropy estimation across various applications, including cryptographic security and anomaly detection.
Claims20
A computer-implemented system for validating cryptographic key randomness in a blockchain environment, the system comprising:
- a memory to store computer executable components;
- a processor to execute the computer executable components stored in the memory, wherein the computer executable components comprise:
- receive an input cryptographic key for use in a blockchain-based security application;
- convert the cryptographic key into a bit sequence for randomness evaluation;
- compute a randomness score for the bit sequence by extracting:
- statistical randomness features, including bit frequency distribution, run-length analysis, Shannon entropy, and min-entropy estimation;
- pattern-based features using a Convolutional Neural Network (CNN) trained to detect structured patterns indicative of low randomness;
- compare the randomness score to a predefined entropy threshold required for blockchain cryptographic security;
- if the randomness score meets or exceeds the threshold, approve the cryptographic key for use in blockchain cryptographic security protocols;
- and
- if the randomness score falls below the threshold, trigger a key regeneration request via a blockchain-based key management system (KMS) or a secure cryptographic module to obtain a new high-entropy key.
A computer-implemented method for validating cryptographic key randomness in a blockchain environment, the method comprising:
- receiving an input cryptographic key for use in a blockchain-based security application;
- converting the cryptographic key into a bit sequence for randomness evaluation;
- computing a randomness score for the bit sequence by:
- extracting statistical randomness features, including bit frequency distribution, run-length analysis, Shannon entropy, and min-entropy estimation;
- and
- extracting pattern-based features using a Convolutional Neural Network (CNN) trained to detect structured patterns indicative of low randomness;
- comparing the randomness score to a predefined entropy threshold required for blockchain cryptographic security;
- if the randomness score meets or exceeds the threshold, approving the cryptographic key for use in the blockchain-based security application;
- and
- if the randomness score falls below the threshold, triggering a key regeneration request via a blockchain-based key management system (KMS) or a secure cryptographic module to obtain a new high-entropy key.
Dependencies are inferred from the claim text.