US 2024/0403572 A1Published application

Cross-lingual document analysis and translation system

AssigneeSchlumberger Technology Corporation

InventorsXiqing Chang, Tianjun Hou, Zheng Zhang

Life of this patent

Estimating the term…

Loading patent…

Abstract

Systems and methods for providing a cross-lingual PDF analysis and translation system specifically designed for the energy industry. The cross-lingual PDF analysis and translation system uses various methods (e.g., tag-based or paragraph-based document extraction and mixing) to create a cross-lingual mixed document dataset by modifying HTML files from multilingual web pages. The cross-lingual mixed document dataset is used to retrain a LayoutXLM model, enabling the LayoutXLM model to further build cross-lingual relations by evaluating the model on various form understanding benchmarks.

Claims20

Dependent claims

A system, comprising:

a network interface configured to retrieve a first plurality of multilingual documents and a second plurality of multilingual documents;
a document generation component configured to generate a cross-lingual mixed document based on the first plurality of multilingual documents;
an artificial intelligence (AI) driven document processing component configured to:
retrain a pre-trained machine learning model to learn cross-lingual relations based on the cross-lingual mixed document;
and
translate the second plurality of multilingual documents based on the cross-lingual relations;
and
one or more task-specific components configured to complement the AI-driven document processing component to implement one or more specific tasks.

5 more elements

Dependencies are inferred from the claim text.