Free AI-901: Microsoft Azure AI Fundamentals Exam Questions and Answers

You plan to use Azure Content Understanding in Foundry Tools to extract information from product photos. For each of the following statements, select Yes if the statement is true. Otherwise, select No. Statement 1: You can define an analyzer with a field schema so that image analysis returns structured fields. Statement 2: You must train a custom machine learning model on labeled photos before an image analyzer can be used. Statement 3: An analyzer for images can return a generated text description of the image as a field.

Answer each statement, then reveal.

1. You can define an analyzer with a field schema so that image analysis returns structured fields.YesNo
2. You must train a custom machine learning model on labeled photos before an image analyzer can be used.YesNo
3. An analyzer for images can return a generated text description of the image as a field.YesNo

A hospital deploys a model that scores each patient's risk of readmission. Clinicians are reluctant to rely on the scores because they cannot tell why a patient was rated high risk. Which action best applies the responsible AI principle of transparency?

• AA. Compare the model's error rates across age groups and report any gaps
• BB. Encrypt the patient records used by the model while they are stored and in transit
• CC. Show clinicians the main factors that contributed to each patient's score
• DD. Assign a named executive to be responsible for the model's clinical outcomes

You are writing a Python app that uses the Azure Speech SDK. You need an object that holds the Speech resource key and region (or endpoint) so that it can be passed to a recognizer or synthesizer. Which class should you use?

• AA. SpeechRecognizer
• BB. AudioConfig
• CC. SpeechSynthesizer
• DD. SpeechConfig

You are building a Python app that uses the Azure Language client library. The app must print the overall label (positive, negative, neutral or mixed) for each hotel review. You have the following code. reviews = ["The room was spotless, but check-in took far too long."] response = client.analyze_sentiment(reviews) for doc in response: print(____) Which expression should you use to complete the code?

• AA. doc.confidence_scores
• BB. doc.entities
• CC. doc.sentiment
• DD. doc.key_phrases

You are building a chat request that sends an image and a question to a deployed multimodal model. The user message contains a content array. Match each content element to its purpose. Each element may be used once, more than once, or not at all. Elements: A. {"type": "text"} entry B. {"type": "image_url"} entry that contains an https address C. {"type": "image_url"} entry that contains a data:image/jpeg;base64 value D. A second system message Targets: Target 1: Carries the question the user asks about the image Target 2: References an image that is publicly reachable on the web Target 3: Embeds a local image file directly in the request

Drag or tap each item onto its target, then reveal.

You are tuning a deployed chat model. You need the model to sample only from the smallest set of candidate tokens whose combined probability reaches a chosen threshold. Which parameter should you configure?

• AA. frequency_penalty
• BB. top_p
• CC. stop sequences
• DD. max_completion_tokens

A company stores long customer support tickets. You need a solution that does both of the following: - Identifies the people, locations and product names mentioned in each ticket. - Produces a short paragraph that condenses each ticket for a manager. Which two text analysis techniques should you use? Each correct answer presents part of the solution. (Choose two.)

Select 2 answers.

• AA. Summarization
• BB. Key phrase extraction
• CC. Sentiment analysis
• DD. Entity detection

You are building an app that uses text to speech to read product names aloud. The synthesized voice mispronounces several product names, and you also need to add short pauses between sentences. What should you use to control this?

• AA. Speech Synthesis Markup Language (SSML) in the synthesis request
• BB. Speaker diarization enabled on the transcription request
• CC. A custom keyword model trained on the product names
• DD. A phrase list added to the speech recognizer configuration

You call an image generation model that is deployed in a Microsoft Foundry project. Which request parameter sets the pixel dimensions of the generated image?

• AA. style
• BB. quality
• CC. n
• DD. size

A law firm must find the company names, dates and monetary amounts in thousands of contracts so that each item can be tagged by category. Which text analysis technique should the firm use?

• AA. Key phrase extraction
• BB. Text summarization
• CC. Entity recognition
• DD. Sentiment analysis