A Search Engine inspired by Natural Language and Human Memory

摘要

In order to improve internet search, a machine learning prototype that combines findings from human memory research with natural language processing will be introduced. Consisting of two parts, one part of the prototype is trained to learn natural language by extracting common sense knowledge from the enormous amount of phrases on the World Wide Web. A machine learning approach that is based on human memory models assists in storing knowledge in an automated database. The second part of the prototype consists of a question-answering system. Once the user enters a query, the database is checked for possible answers. These will be provided in case there is an overlap between the user's query and the stored information.