Microsoft AI-102 - Designing and Implementing a Microsoft Azure AI Solution

Removing all the chit-chat question and answer pairs from the project will not ensure that the chatbot provides formal responses to spurious questions. It will only make the chatbot unable to handle any chit-chat scenarios, which may result in a poor user experience and a loss of engagement. Instead, you should choose a chit-chat personality that matches the tone and style of your chatbot, such as Professional or Caring. You can also edit the chit-chat questions and answers to suit your specific needs, or add new ones that are not in the predefined data set12. This way, you can ensure that the chatbot responds appropriately to spurious questions, while still maintaining a conversational and engaging interaction with the user.

Comprehensive Detailed ExplanationIn the provided LUDown file snippet:

Intents represent the purpose or action the user wants to perform.

Examples: ## Confirm, ## ExtractName, ## Logout, ## SelectItem, ## SelectNone.

These are all intent definitions.

So, SelectItem is an intent.

Utterances are the example phrases that users might say to invoke a particular intent.

Under ## SelectItem, one example utterance is:

choose {@DirectionalReference=top right}

This is an example phrase that maps to the SelectItem intent and contains the DirectionalReference entity.

Therefore, Choose {@DirectionalReference=top right} is an utterance.

Domain: Predefined category grouping multiple intents (not applicable here).

Entity: Extracted data within utterances (e.g., @DirectionalReference=top right or @personName).

The entire phrase "choose {@DirectionalReference=top right}" is not an entity but an utterance containing an entity.

Why not the other options?

SelectItem → an intent

Choose {@DirectionalReference=top right} → an utterance

Correct Answers

LUDown file format for LUIS

LUIS concepts: intents, utterances, and entities

Microsoft References

To create a resource that can generate fictional stories from prompts, you need an Azure OpenAI account (--kind OpenAI).

To protect data with a customer-managed key (CMK), you must pass the key information (Key Vault URI, key name, key version) via the --encryption parameter when creating the Cognitive Services account. The JSON shown (with keyName, keyVersion, and keyVaultUri) is exactly the payload expected by --encryption.

A complete example pattern:

az cognitiveservices account create \

-n myresource -g myResourceGroup -l westeurope \

--kind OpenAI --sku S \

--assign-identity \

--encryption '{

"keyName":"keyname",

"keyVersion":"secretVersion",

"keyVaultUri":"https:// .vault.azure.net/"

}'

(Adding --assign-identity allows the service to access the Key Vault key; the hotspot specifically asks for the parameter that takes the JSON block and the resource kind, which are --encryption and OpenAI.)

Microsoft References

Azure CLI: Create Cognitive Services account with customer-managed keys using --encryption.https://learn.microsoft.com/azure/ai-services/cognitive-services-encryption-keys#configure-customer-managed-keys

Azure Cognitive Services (Azure AI services) CLI parameters including --kind OpenAI.https://learn.microsoft.com/azure/ai-services/cognitive-services-management-api-cli#create-a-cognitive-services-account

The code creates a SpeechConfig and an AudioOutputConfig with filename="path/to/file.wav", then constructs a SpeechSynthesizer and calls speak_text_async(input).

Existing file behavior: AudioOutputConfig with a filename directs synthesis output to that file path; the documentation emphasizes that it specifies the output file and that the parent directory must already exist. It does not require a unique filename (i.e., it does not fail purely because a file exists). In practice, the SDK writes to the specified path. Therefore, the statement claiming failure if a file already exists is No.

Voice source: The filename in AudioOutputConfig is output, not input. To choose a voice, you set speech_config.speech_synthesis_voice_name or use SSML; the WAV file isn’t “sampled” to create a voice. Hence this statement is No. Microsoft Learn

File generation: Using AudioOutputConfig(filename=...) with SpeechSynthesizer and speak_text_async(...) writes synthesized audio to that file. Many samples show the same pattern (often with .get() to wait for completion). So the function’s purpose is to generate an audio file from the input text—Yes.

Note: In production, call .get() (or await) on speak_text_async(...) so the program waits for synthesis to complete before exiting. Microsoft Learn

References (Microsoft Azure AI Speech)

How to synthesize speech from text (create AudioOutputConfig to write to a file; synthesize with speak_text_async).

Python AudioConfig class (audio output can be a speaker, audio file (WAV), or stream). Microsoft Learn

.NET AudioConfig.FromWavFileOutput (specifies output file; parent directory must already exist). Microsoft Learn

Azure Q&A example (Python: speak_text_async(...).get() usage). Microsoft Learn

 The code will perform face recognition → No

 The code will list tags and their associated confidence → Yes

 The code will read a file from the local file system → Yes

The given method uses the Computer Vision client library:

It specifies the features:

List features = new List()

{

VisualFeatureTypes.Description,

VisualFeatureTypes.Tags,

};

That means it will return descriptions (captions) and tags from the image.

The image is opened from the local file system:

using (Stream imageStream = File.OpenRead(localImage))

The analysis call:

ImageAnalysis results = await client.AnalyzeImageInStreamAsync(imageStream, features);

It prints captions and tags with their confidence values.

“The code will perform face recognition.”

No.

The requested features are Description and Tags, not Faces.

Face recognition would require VisualFeatureTypes.Faces.

“The code will list tags and their associated confidence.”

Yes.

The foreach (var tag in results.Tags) loop outputs tag name and confidence.

“The code will read a file from the local file system.”

Yes.

It uses File.OpenRead(localImage) which reads from disk.

Statement Analysis

The code will perform face recognition → No

The code will list tags and their associated confidence → Yes

The code will read a file from the local file system → Yes

Final Answer

Analyze an image with the Computer Vision API

VisualFeatureTypes enum reference

Microsoft References

In the Microsoft Bot Framework SDK, bots communicate using activities (messages, events, conversation updates, etc.).

To respond to events (like conversation started, member added, or custom event activities), you use an ActivityHandler.

The ActivityHandler class provides methods you can override, such as OnMessageActivityAsync, OnMembersAddedAsync, or custom event handlers, to send custom text responses.

Other options:

Adaptive card (A) → used for rich UI responses, not for handling events.

Dialog (C) → used for managing conversation flow, not event handling.

Skill (D) → used to integrate another bot’s functionality, not for handling custom events.

Correct Answer: B

The system must analyze multiple sensor data points (rotation speed, angle, temperature, pressure) and generate alerts on anomalies.

Application Insights → used for monitoring applications, not IoT sensor data.

Metric alerts in Azure Monitor → works with single metrics but does not handle complex correlations.

Univariate Anomaly Detection → only works on a single time series (e.g., just temperature).

Multivariate Anomaly Detection → specifically built to detect anomalies across multiple correlated variables (sensor data).

Thus, the correct choice is Multivariate Anomaly Detection.

The Text Moderation API of Azure Content Moderator is specifically used for scanning text to detect potentially inappropriate or undesired content.

Key features include:

Adult classification score → Indicates the likelihood that the text contains adult content.

Racy classification score → Indicates the likelihood that the text contains sexually suggestive but not explicit content.

PII detection (personal data) is part of text moderation but is not always required when focusing strictly on enabling content moderation for chatbot text safety.

OCR is for extracting text from images, not for moderating chatbot conversations.

Text classification in this context refers to general machine learning categorization (not part of Content Moderator’s moderation API).

Thus, the two service responses directly tied to text moderation safety are:

Adult classification score

Racy classification score

Correct Answer for Q185: A and E

Azure Content Moderator text moderation overview

Microsoft Reference

The requirements are:

Generate tags for uploaded images → This is an image analysis task.

Tags must be returned in the user’s preferred language → Must support localization of tags.

Languages required: English, French, and Spanish → Must support multiple languages.

Minimize development effort → Should use a prebuilt service rather than training a custom model.

Option Analysis:

A. Custom Vision image classification

Custom Vision requires training a model with your own dataset.

It outputs classification labels, but these are defined in English by the developer, and localization would require extra translation work.

More development effort, not ideal here.

B. Content Moderator Image Moderation

Used to detect adult/racy content, profanity, or unsafe imagery.

Does not generate descriptive image tags.

Not suitable.

C. Custom Translator

Used to train translation models for text.

Does not analyze or generate image tags.

Not applicable here.

D. Computer Vision Image Analysis

Provides prebuilt image tagging and description features.

Automatically generates descriptive tags for uploaded images.

Supports multiple languages (including English, French, and Spanish) for returning tags and descriptions.

Minimal development effort: simply call the Image Analysis API with the desired language parameter.

This exactly meets the requirements.

Correct Answer: D. Computer Vision Image Analysis

Azure AI Vision Image Analysis – Tags and Descriptions

Azure AI Vision supported languages

Quickstart: Analyze an image with the Computer Vision REST API

Microsoft References

Azure AI Content Safety supports custom blocklists—dictionaries of terms/phrases you define to detect obscure or domain-specific offensive terms in user content. Creating a blocklist is the minimal-effort, built-in way to meet this requirement; you don’t need to build a classifier.

References

Create and use blocklists in Azure AI Content Safety.https://le arn.microsoft.com/azure/ai-services/content-safety/concepts/blocklists

Text moderation overview (using built-in classifiers with optional blocklists).https://learn.microsoft.com/azure/ai-services/content-safety/concepts/text-moderation