Graph indexing, storage, and retrieval with cognitive memory

摘要

PROBLEM TO BE SOLVED: To provide a non-transitory computer-readable medium for graph or network encoding and indexing and identification of a substructure match or similar in graph data. The cognitive memory computes the local structure of a graph in addition to contraction and calculation of graph properties to build an index data structure called a cognitive signature for searches based on properties, similarities, or structures or substructures. To summarize, the graph is encoded by a generalized combinatorial map and a new quantum-inspired Q-Hashing algorithm. The method makes the system ideal for semantic indexing of large graph datasets by using the properties of the graph's cognitive signature to correlate with the properties of the underlying data. [Selection diagram] Figure 1