US 11,790,245 B2
Cognitive machine learning for semantic network
Khursheed Sheikh, Los Angeles, CA (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Jan. 22, 2019, as Appl. No. 16/253,437.
Prior Publication US 2020/0234180 A1, Jul. 23, 2020
Int. Cl. G06F 16/2458 (2019.01); G06N 5/02 (2023.01)
CPC G06N 5/02 (2013.01) [G06F 16/2458 (2019.01)] 19 Claims
OG exemplary drawing
 
1. A computer program product comprising:
a computer readable storage medium readable by one or more processor and storing instructions for execution by the one or more processor for performing a method comprising:
generating a semantic network cell that corresponds to a component of a semantic expression in a semantic network, wherein the semantic network includes a plurality of semantic network cells, and wherein each semantic network cell has attributes of a weight, an access count, and a latest time of access;
operating the semantic network as a knowledgebase servicing searches in dependence on a machine learning process that emulates reinforcing of the semantic network cell by access and decaying of the semantic network cell over time in dependence on semantic network cell weight rules, wherein the operating includes obtaining search results for iterations of semantic queries of the semantic network received from a search engine, and identifying semantic network cells of the semantic network that are accessed as a result of respective iterations of semantic queries, wherein the accessed semantic network cells that are accessed as a result of the respective iterations of semantic queries satisfy conditions of the respective iterations of the semantic queries; and
strengthening weights of the identified semantic network cells of the semantic network that are accessed as a result of the respective iterations of semantic queries in dependence on strength factors of the semantic network cell weight rules, wherein the strengthening is in dependence on the access count, and latest time of access attributes of respective ones of the identified semantic network cells that are accessed as a result of the respective iterations of semantic queries.