|
| 1 | +### YamlMime:ModuleUnit |
| 2 | +uid: learn.wwl.extract-data-with-document-intelligence.knowledge-check |
| 3 | +title: Module assessment |
| 4 | +metadata: |
| 5 | + title: Module assessment |
| 6 | + description: Knowledge check for extracting data with Azure Document Intelligence. |
| 7 | + ms.date: 03/11/2026 |
| 8 | + author: ivorb |
| 9 | + ms.author: berryivor |
| 10 | + ms.topic: unit |
| 11 | + ai-usage: ai-assisted |
| 12 | + module_assessment: true |
| 13 | +durationInMinutes: 3 |
| 14 | +content: | |
| 15 | + [!include[](includes/7-knowledge-check.md)] |
| 16 | +quiz: |
| 17 | + title: "Check your knowledge" |
| 18 | + questions: |
| 19 | + - content: "You need to extract text and table structure from a set of documents that have varying formats. You don't need to identify specific labeled fields. Which Document Intelligence model should you use?" |
| 20 | + choices: |
| 21 | + - content: "The read model." |
| 22 | + isCorrect: false |
| 23 | + explanation: "Incorrect. The read model extracts text and languages but doesn't extract tables or structural information." |
| 24 | + - content: "The layout model." |
| 25 | + isCorrect: true |
| 26 | + explanation: "Correct. The layout model extracts text, tables, selection marks, and document structure information, making it ideal for documents with varying formats where you need structural data." |
| 27 | + - content: "The invoice model." |
| 28 | + isCorrect: false |
| 29 | + explanation: "Incorrect. The invoice model is a prebuilt model designed specifically for invoices, not general documents with varying formats." |
| 30 | + - content: "You're building a custom model in Azure Document Intelligence. What training artifacts are required when training with the REST API?" |
| 31 | + choices: |
| 32 | + - content: "Only the sample form documents in a blob container." |
| 33 | + isCorrect: false |
| 34 | + explanation: "Incorrect. In addition to sample forms, you need JSON files that describe the fields, labels, and OCR data for each form." |
| 35 | + - content: "Sample forms along with ocr.json, labels.json, and fields.json files in a blob container." |
| 36 | + isCorrect: true |
| 37 | + explanation: "Correct. You need an ocr.json file for each sample form, a labels.json file for each form mapping fields to locations, and a single fields.json file describing the fields to extract." |
| 38 | + - content: "A minimum of 100 labeled forms and a trained classifier." |
| 39 | + isCorrect: false |
| 40 | + explanation: "Incorrect. You can train a custom model with as few as five to six sample forms. A classifier is optional and separate from the extraction model." |
| 41 | + - content: "A company processes both invoices and receipts. They want a single endpoint that routes each document to the correct extraction model. What should they use?" |
| 42 | + choices: |
| 43 | + - content: "A custom neural model." |
| 44 | + isCorrect: false |
| 45 | + explanation: "Incorrect. A custom neural model extracts data from one type of document. It doesn't automatically route between different document types." |
| 46 | + - content: "A prebuilt read model." |
| 47 | + isCorrect: false |
| 48 | + explanation: "Incorrect. The read model extracts text but doesn't classify documents or extract domain-specific fields." |
| 49 | + - content: "A composed model or a custom classifier paired with extraction models." |
| 50 | + isCorrect: true |
| 51 | + explanation: "Correct. A composed model combines multiple custom models into a single endpoint and classifies each document to the appropriate component model. Alternatively, a custom classifier can identify the document type before routing to the correct extraction model." |
| 52 | + |
0 commit comments