Document Information Extraction
An AI task that locates and extracts predefined fields, such as invoice numbers, contract amounts and dates, from unstructured documents. Evaluation metrics include field-level exact match, recall and end-to-end accuracy.
Timeline (last 90 days)
When enterprises deploy AI for document information extraction, they upload large numbers of long-tail complex documents, and the diversity of these documents can even cause the most advanced models to make errors.
Document information extraction tasks typically require a model to locate and extract predefined fields, such as invoice numbers, contract amounts, and dates, from unstructured documents.
When evaluating document information extraction tasks, commonly used metrics include field-level exact match, recall, and end-to-end accuracy.
All Facts (3)
When enterprises deploy AI for document information extraction, they upload large numbers of long-tail complex documents, and the diversity of these documents can even cause the most advanced models to make errors.
50%UnverifiedDocument information extraction tasks typically require a model to locate and extract predefined fields, such as invoice numbers, contract amounts, and dates, from unstructured documents.
50%UnverifiedWhen evaluating document information extraction tasks, commonly used metrics include field-level exact match, recall, and end-to-end accuracy.
50%