Machine-learning document categorization enables turnkey email, ticket, and report routing systems from expiring IBM IP.
The method receives a document, extracts features from its content, applies a learned classifier, and assigns one or more categories based on the classifier’s output. Its technical significance is that it turns feature extraction plus a classifier into an automated multi-category labeling pipeline, replacing manual rules at scale. Previously, deploying IBM’s trained approach in production required licensing or rights to the patented method/apparatus rather than using it freely.
Build automated document triage APIs for customer support, legal intake, and compliance reporting without ongoing patent royalties.
A method and apparatus for categorizing documents using machine learning, including receiving a document, extracting features from the document, applying a learned classifier to the extracted features, and assigning the document to one or more categories based on outputs of the classifier.
Post a problem you're solving with this technology and connect with founders, investors, and researchers.