Methods, apparatuses and systems for annotation of text documents

Through the Alpine system, field experts can independently design and execute labeling projects, solving the problem that traditional tools are difficult to quickly generate high-quality labeling data sets, and achieving rapid generation and high-quality output of labeling data sets, which are highly adaptable.

JP2025037908AActive Publication Date: 2025-03-18PAREXEL INTERNATIONAL LLC
4 Cites 0 Cited by

Patent Information

Application Number
JP2024202391
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2019-03-11
Filing Date
2024-11-20
Publication Date
2025-03-18
Estimated Expiration
2040-03-11

AI Technical Summary

Technical Problem

It is difficult for the prior art to quickly generate high-quality annotation data sets, especially in industrial applications. Traditional annotation tools are independent of the label allocation process, ignore the coordinated investigation steps, and are difficult to deal with changes in annotation schemes and guidelines.

Method used

The Alpine system was designed, an integrated GUI based on tools that allowed domain experts to independently design and execute annotation projects. Alpine supports a non-single collaborative workflow, allowing multiple personal annotators to investigate documents, create and adjust labels and guides, and integrate these discussions into the label manual, adjust labels and labels in real-time, research-generated label documents.

Benefits of technology

The Alpine system enables field experts to work directly with project managers and clients independently of the engineering team to define customized annotation projects, improving the quality and speed of the annotation data set, being highly adaptable and able to respond quickly to business needs.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

To provide methods and apparatuses to facilitate annotation projects to extract structured information from free-form text using natural language processing (NLP) techniques.SOLUTION: Annotators explore text documents via automated pre-annotation functions, flexibly formulate annotation schemes and guidelines, annotate text, and adjust annotation labels, schemes and guidelines as a project evolves. NLP models are readily trained on iterative annotations of sample documents by experts in an active learning workflow. The trained models are then employed to automatically annotate a larger body of documents in a project dataset. Experts can readily develop an annotation project for a specific use-case or business question.SELECTED DRAWING: Figure 12
Need to check novelty before this filing date? Find Prior Art