Machine Learning | Functionality Overview

Lemmana Content Services uses the latest in Computer Vision, Natural Language Processing, Machine Learning and AI approaches to simplify and reduce the cost of document processing for your business and drive Digital Transformation.

Initially, a set of known documents are sorted into specific categories or classes by your knowledge workers. The classes align to your business sorting needs i.e. Invoices, Statements, Payment Letters, KYC etc. Specific key data is also identified by the workers within each of these classes, such as Reference, Amount, Date etc. Lemmana Content Services then automatically learns from these examples and creates a custom trained model. This model is then applied when an unknown document is passed through Lemmana Content Services. The documents are intelligently analysed by classifying and capturing the required data. This document can then be routed to the correct workflow or storage location with DLX DM5.

The custom model that is created is dependent upon the samples the Lemmana Content Services is given to analyse. Therefore, it can improve or degrade based upon the variety and quantity of document samples provided for training. To aid in assuring you get optimal results, Lemmana Content Services can run tests on the samples it holds to give estimations of expected accuracy. These results are displayed graphically within the Lemmana Content Services web UI. This allows the user to identify which document classes and metadata may need more or less samples/training.