Cognitive Memory Graph Indexing, Storage, and Retrieval Techniques

摘要

The present disclosure provides fast approximation and accurate hierarchical network storage and retrieval systems and methods for encoding and indexing graphs or networks and identifying substructure matching or similarity in graph data. New quantum intuitive Q-hashing and generalized combinatorial maps that summarize the local structure of graphs based on reduction and graph property calculations to build index data structures called cognitive signatures for attribute-based, similarity-based, or structured or substructure-based searches. The algorithm encodes the graph. The systems and methods of the present invention are suitable for storing and indexing all or part or substructures or analogues of graphs, as well as dynamically changing graphs such as traffic graphs or flow and video sequences of graphs. This system and method has the advantage that the attributes of the cognitive signature of the graph can be used for association with the attributes of the underlying data, which makes the system ideal for semantic indexing of large graph data sets.