AI/ML Discovery

Natural Language Processing & Pattern Discovery

One of the primary features of this portal is that information from across multiple performers can be readily processed by AI machine learning.

First is natural language process (NLP). The words and phrases from the documents uploaded into this system are parsed and stored in the database, with reference to document ID, paragraph ID, sentence ID and the word or phrase of interest. Generally those of interest have corresponding quantum ontology terms, referred to here as Qlabels. Generally the words and phrases of interest are those contained in the Quantum Glossary. Total counts are tallied for each work and Quantum Phrase.

AI machine learning uses the above data to a) cluster and b) categorize. Clustering is unsupervised, meaning the machine learning occurs with labels from the data. This is useful in discovering patterns between authors (SMEs), countries, institutions, funding sources and many other aspects. Categorization is achieved by supervised machine learning meaning labels from each document are used in the machine learning training sessions. Information from within the documents, such as keywords or Quantum Phrases, or extraneous data, can both be used as the labels. The labeled dataset is split into the training set and the testing set that is used to test the degree of accuracy of the learned model. Almost anything can be modeled using machine learning, making AI machine learning potentially the highest-impact feature of this portal.