Artificial intelligence (AI) based innovation data processing system

An artificial intelligence (AI) based innovation identification method comprising:

• receiving at least one query word related to a category pertaining to an innovation identification process, wherein different stages of innovations can be identified within the category;
• search a plurality of data sources for documents of different types using the at least one query word as a search query;
• extract a plurality of documents related to the at least one query word from the plurality of data sources, the plurality of documents including at least textual data related to one or more innovations, benefits of the innovations and suppliers of the innovations;
• associating each of the plurality of documents with one or more of the different stages based at least on a type of the document;
• identifying entities from the textual data of the plurality of documents wherein the entities are associated with the innovation identification within the category;
• selecting at least portions of the plurality of documents based on material selection criteria pertaining to the entities for further processing;
• generating, via dependency parsing, labeled syntactic dependencies between words in the portions of the plurality of documents;
• obtaining, via providing the labeled syntactic dependencies to a trained classifier, a subset of one or more of the entities corresponding to the innovations from the portions of the plurality of documents;
• generating one or more knowledge graphs including the subset of one or more entities and relationships between the subset of one or more entities;
• determining trends in respective frequencies of occurrences of the subset of one or more entities pertaining to the innovations in the portions of the plurality of documents over a specified time period;
• ranking the subset of one or more entities based on the frequencies of the subset of one or more entities in the portions of the plurality of documents;
• identifying as an innovation within the category, at least one of the ranked subset of one or more entities that has a lower frequency of occurrences in an initial time period of the specified time period;
• and
• classify the innovation into one of the different stages of innovation based at least on the association of each of the plurality of documents with the one or more stages.

A non-transitory processor-readable storage medium comprising machine-readable instructions that cause a processor to:

• receive at least one query word related to innovation identification within a category, wherein different stages of innovations can be identified within the category;
• search a plurality of data sources for documents of different types using the at least one query word as a search query;
• extract a plurality of the documents related to the at least one query word from at least one data source of the plurality of data sources, the plurality of documents including at least textual data related to the innovation identification within the category;
• associate each of the plurality of documents with one or more of the different stages of innovation based at least on a type of the document;
• identify entities in the textual data of the plurality of documents wherein the entities are associated with the innovation identification within the category;
• select portions of the plurality of documents that satisfy material selection criteria pertaining to the entities for further processing;
• generate, via dependency parsing, labeled syntactic dependencies between words in sentences of the selected portions of the plurality of documents;
• identify, via providing the labeled syntactic dependencies to a trained classifier, a subset of one or more of the entities corresponding to the innovation identification from the selected portions of the plurality of documents;
• generate a knowledge graph including the subset one or more entities and relationships between the subset one or more entities;
• determine trends in respective frequencies of occurrences of the subset of one or more entities in the portions of the plurality of documents over a specified time period;
• rank the subset of one or more entities based on the frequencies of the subset of one or more entities in the portions of the plurality of documents;
• and
• output the subset of one or more entities, using the knowledge graph and the rankings;
• identify as an innovation within the category, at least one of the ranked subset of one or more entities that has a lower frequency of occurrences in an initial time period of the specified time period;
• and
• classify the innovation into one of the different stages of innovation based at least on the association of each of the plurality of documents with the one or more stages.