Building a QnA Solution

A knowledge base is what?

• Question and answer pairs that can be queried using natural language input

• Question and answer pairs that can be queried using exact language input

• A list of facts that can be queried using natural language input

• A list of questions that can be queried using exact language input

What are two ways you can populate a knowledge base with existing question and answer pairs?

• Import questions and answers from existing web pages or documents.

• Import answers from an existing web pages or documents and use AI to create the questions.

• Add pre-defined “chit-chat” pairs that include common conversational questions and responses.

• You cannot populate a knowledge base with existing question and answer pairs.

Can you create a QnA Maker solution without using the QnA Maker portal?

• Yes, you can use the REST API or SDK to write code that defines, trains, and publishes the knowledge base.

• Yes, you can use machine learning to write code that defines, trains, and publishes the knowledge base.

• No, you must use the QnA Maker to create a QnA Maker solution.

• No, but you can create a knowledge base using the Knowledge SDK.

Both language understanding and QnA Maker Knowledge base are forms of language models, but there are some differences. What action does the QnA Maker enable that the Language Understanding service does not?

• User submits an utterance, expecting an appropriate response or action.

• Response indicates the most likely intent and referenced entities.

• Response is a static answer to a known question.

• None

What service is used to create a bot from your knowledge base?

• QnA Maker

• Language Understanding

• Text Analytics

• Language Analytics

What are the primary differences between the QnA Maker and Language Understanding?

What is the difference between Implicit and Explicit feedback?

In the QnA client interface, if the request body contains a question, what is included in the response?

Using the Q&A maker, can you train and test a natural language model before publishing it for use in an application or bot?

Can QnA maker handle multi-turn conversations?