A platform that accelerates the acquisition of knowledge through semantic search and knowledge path by modelling organization knowldge through internal experts and internal documents.

Semantic search highlight few important areas relevant to an entity (instead of just a link) in normal search. Those areas come from nodes of a knowledge graph relevant to the entity.

In order to map keywords to a node in a knowledge graph, a simple index-based approach will need to be extended. Since query keywords will need to point to any nodes, these keywords will be transformed into word embeddings first, before performing any lookup. Word Embeddings is a matrix of numbers representing the keywords. Producing Word Embeddings is a part of Natural Language Processing.

To understand more about semantic search, we construct a sample user search simulation. Download the whitepaper below to learn more.