ML Model and Foundation Model (FM) Development A free question from our original MLA-C02 practice exam. Try it, then open the answer to see why each option is right or wrong.
A dental chain receives scanned insurance claim forms from patients by mail and fax. Each form has labeled fields such as patient name, member ID and procedure codes, and many forms include tables listing each treatment and its charge.
The chain has no machine learning team. It needs the field values and the tables pulled out as structured data with the least custom development.
Which AWS service should the chain use?
Choose one answer.
AA. Amazon Rekognition, using its text detection API to find the words on each page and then pairing labels with values in custom code.
BB. Amazon Comprehend, using its entity recognition on the scanned forms after they are saved as text files.
CC. Amazon Textract, calling AnalyzeDocument with the FORMS and TABLES feature types to return key-value pairs and table cells.
DD. Train a custom convolutional neural network in SageMaker AI on the scanned forms, labeling every field by hand first.
Show the answer and explanation
Answer: C
A. Wrong: Rekognition detects objects, faces and some text in images, but it does not return form key-value pairs or table structure. The chain would have to build the extraction itself.
B. Wrong: Comprehend analyzes text that is already available, and it does not read scanned pages or extract table cells. The forms need a document extraction service first.
C. Right: Correct. Textract reads scanned documents and, with the FORMS and TABLES feature types, returns labeled fields and table cells as structured data. No custom model is needed. Learn more: What is Amazon Textract? (opens in a new tab)
D. Wrong: A custom model needs many labeled forms and ML expertise, which the chain does not have. A managed document service already covers this task with less effort.
Explanation
Amazon Textract extracts text, form fields and tables from scanned documents, and AnalyzeDocument with FORMS and TABLES returns them as structured data. Rekognition and Comprehend do not parse forms and tables, and a custom model is more work than this managed service needs.
Want more on AWS AI services? The 5 full-length practice sets have 5 more AWS AI services questions, 325 in all, none of them in the free exam. One payment of $14.99, no subscription. See the practice sets →