Microsoft Azure AI Engineer Associate AI-102 (AI-102) — Questions 826900

988 questions total · 14pages · All types, answers revealed

Page 11

Page 12 of 14

Page 13
826
MCQmedium

You are designing a solution that uses Azure OpenAI Service to generate product descriptions based on product attributes (name, category, features). The solution must: - Use a GPT-4 model deployed in the West US region. - Implement content filtering to block inappropriate content. - Handle up to 100 requests per second. - Minimize latency. - Use managed identity for authentication. What should you include in the design?

A.Deploy multiple GPT-4 models across different regions to reduce latency.
B.Deploy a single GPT-4 model with sufficient capacity (e.g., 100K TPM). Enable content filtering. Use a system-assigned managed identity for authentication.
C.Use API key authentication stored in Azure Key Vault. Deploy two GPT-4 models to load balance requests.
D.Use Azure AI Content Safety in addition to Azure OpenAI to filter content.
AnswerB

Meets all requirements with minimal complexity.

Why this answer

Option B is correct because deploying a single GPT-4 model with sufficient capacity (e.g., 100K TPM) ensures the solution can handle up to 100 requests per second while minimizing latency by avoiding cross-region calls. Enabling content filtering directly on the Azure OpenAI deployment blocks inappropriate content without additional services, and using a system-assigned managed identity provides secure, keyless authentication that aligns with Azure best practices.

Exam trap

The trap here is that candidates often overcomplicate the solution by adding unnecessary redundancy (multiple models or regions) or extra services (Azure AI Content Safety), when the built-in capabilities of Azure OpenAI—content filtering, managed identity, and sufficient TPM—directly satisfy all requirements with minimal latency.

How to eliminate wrong answers

Option A is wrong because deploying multiple GPT-4 models across different regions would increase latency due to cross-region network hops and does not address the requirement to minimize latency, as the model must be in West US. Option C is wrong because using API key authentication stored in Azure Key Vault violates the requirement to use managed identity for authentication, and deploying two GPT-4 models for load balancing is unnecessary when a single model with sufficient TPM capacity can handle 100 requests per second. Option D is wrong because Azure AI Content Safety is an additional service that adds latency and complexity; Azure OpenAI’s built-in content filtering already meets the requirement to block inappropriate content without needing an extra component.

827
MCQeasy

A financial services company uses Azure AI Language to analyze customer support transcripts. They want to identify the main topics discussed in each conversation and generate a summary of the key points. The solution must minimize development effort and use prebuilt functionality. You need to recommend the appropriate Azure AI Language features. What should you use?

A.Custom Named Entity Recognition (NER) and conversation summarization.
B.Key phrase extraction and conversation summarization.
C.Entity linking and conversation summarization.
D.Sentiment analysis and key phrase extraction.
AnswerB

Key phrase extraction identifies topics, and conversation summarization generates summaries.

Why this answer

Conversation summarization is a prebuilt feature that generates summaries of conversations, and key phrase extraction identifies main topics. Both are available in Azure AI Language without custom training. Option C combines these correctly.

Option A is wrong because custom NER requires labeling. Option B is wrong because sentiment analysis does not summarize. Option D is wrong because entity linking is not for summarization.

828
MCQhard

Your team is using Azure AI Search to index a large collection of technical manuals. Users report that searches for 'disk failure' do not return relevant results because the manuals use terms like 'hard drive crash'. Which feature should you implement to improve recall?

A.Apply a filter
B.Configure a scoring profile
C.Enable semantic search
D.Add a synonym map to the index
AnswerD

Synonyms map equivalent terms to improve recall.

Why this answer

Option A is correct because a custom synonym map can define equivalent terms like 'disk' and 'hard drive'. Option B is wrong because semantic ranking re-ranks results but does not add synonyms. Option C is wrong because scoring profiles boost by fields, not synonyms.

Option D is wrong because filters reduce results but do not expand queries.

829
MCQmedium

You are using Azure AI Language to analyze customer reviews. You need to identify specific aspects (e.g., 'price', 'service') and their associated sentiment. Which feature should you use?

A.Opinion Mining
B.Sentiment Analysis
C.Aspect-based sentiment analysis
D.Key Phrase Extraction
AnswerC

Aspect-based sentiment analysis identifies aspects and their sentiment.

Why this answer

Option D is correct because Aspect-based sentiment analysis (ABSA) identifies aspects and their sentiment. Option A is wrong because standard sentiment analysis gives overall sentiment. Option B is wrong because key phrase extraction identifies phrases but not sentiment.

Option C is wrong because opinion mining is a subset of ABSA; ABSA is the correct feature name.

830
MCQeasy

A company wants to moderate user-generated images for adult content. Which Azure AI Vision feature should they use?

A.Custom Vision with a custom adult classifier
B.Face API
C.Analyze Image API with moderation categories
D.OCR
AnswerC

The Analyze Image API can detect adult, racy, and gory content.

Why this answer

Option B is correct because the Analyze Image API includes adult content moderation. Option A is wrong because OCR is for text extraction. Option C is wrong because Custom Vision can be trained for moderation but is not the simplest solution.

Option D is wrong because the Face API is for face detection and recognition.

831
MCQeasy

Refer to the exhibit. You have a skillset with two skills. You run the indexer and find that the output field 'organizations' is empty for documents that clearly contain organization names. The 'keyPhrases' output is populated correctly. What is the most likely cause of the issue?

A.The skill's 'name' property is set to '#1', which is invalid.
B.The skill is not configured with a 'defaultLanguageCode' and the documents are not in English.
C.The 'categories' property is misspelled; it should be 'entityCategories'.
D.The input source '/document/content' is incorrect; it should be '/document/text'.
AnswerB

EntityRecognitionSkill needs a language hint to perform correctly for non-English languages.

Why this answer

The 'keyPhrases' output is populated correctly, indicating the text extraction and overall pipeline are functional. The Entity Recognition skill requires a 'defaultLanguageCode' to correctly identify entities; if it is not set and the documents are not in English, the skill may fail to extract organizations, resulting in an empty 'organizations' field. This is a known behavior where the skill defaults to English and cannot process other languages without explicit configuration.

Exam trap

The trap here is that candidates assume the 'keyPhrases' skill working correctly implies all skills are fine, overlooking that Entity Recognition is language-dependent and requires explicit 'defaultLanguageCode' configuration, while 'keyPhrases' is more robust across languages.

How to eliminate wrong answers

Option A is wrong because the skill's 'name' property being set to '#1' is not invalid; skill names can include special characters and are only used for identification within the skillset, not for functionality. Option C is wrong because the 'categories' property is correctly spelled for the Entity Recognition skill (it uses 'categories' to specify which entity types to extract, such as 'organization'); there is no property named 'entityCategories'. Option D is wrong because '/document/content' is the correct default input path for content extracted from most data sources (e.g., Azure Blob Storage), while '/document/text' is not a standard field in the enriched document tree unless explicitly mapped.

832
Multi-Selectmedium

Which TWO Azure AI services can be used to perform optical character recognition (OCR) on images? (Choose two.)

Select 2 answers
A.Azure AI Document Intelligence Read model
B.Azure Video Indexer
C.Azure AI Custom Vision
D.Azure AI Face API
E.Azure AI Vision OCR (Read API)
AnswersA, E

Document Intelligence offers a Read model for OCR with layout preservation.

Why this answer

Azure AI Vision OCR (Read API) and Azure AI Document Intelligence Read model both provide OCR capabilities. Option C is wrong because Custom Vision is for image classification/object detection. Option D is wrong because Video Indexer is for video analysis.

Option E is wrong because Face API is for face detection.

833
MCQhard

You are implementing an agentic solution using Azure AI Agent Service with multiple agents that need to collaborate. Each agent has access to different knowledge bases. You want to ensure that the agents can share context and hand off tasks to each other seamlessly. Which architecture should you use?

A.Create a single monolithic agent that includes all knowledge bases
B.Deploy each agent independently and configure them to call each other via HTTP
C.Use a supervisor agent that delegates to specialized agents, with a shared context store in Azure Cosmos DB
D.Chain the agents sequentially, passing output from one to the next
AnswerC

Supervisor pattern with shared context enables seamless handoff.

Why this answer

Option A is correct because a supervisor agent with a shared context store allows coordination and handoff. Option B is incorrect because separate independent agents lack coordination. Option C is incorrect because sequential chaining limits flexibility.

Option D is incorrect because a single monolithic agent defeats the purpose of specialization.

834
MCQhard

A healthcare company is using Azure AI Document Intelligence to extract patient data from forms. They need to ensure that all extracted data is encrypted at rest using a customer-managed key (CMK) and that the service endpoint is restricted to a specific virtual network. Which combination of steps should they take?

A.Use a service endpoint and configure a managed identity
B.Disable public network access and enable CMK via Azure Key Vault
C.Configure IP firewall rules and enable CMK via Azure Key Vault
D.Create a private endpoint and associate a customer-managed key in the resource encryption settings
AnswerD

Private endpoint secures network traffic; CMK encryption is configured in resource settings with Key Vault.

Why this answer

Option D is correct because it combines a private endpoint (which restricts the service endpoint to a specific virtual network by providing a private IP address within that VNet, eliminating public internet exposure) with a customer-managed key (CMK) in the resource encryption settings, which ensures data at rest is encrypted using a key stored in Azure Key Vault that the customer controls. This directly meets both requirements: network isolation via private endpoint and CMK-based encryption at rest.

Exam trap

The trap here is that candidates often confuse 'service endpoint' or 'IP firewall rules' with 'private endpoint' for VNet-specific access, but only a private endpoint provides a fully private IP within the VNet and meets the 'restricted to a specific virtual network' requirement, while the other options either allow public exposure or do not enforce VNet-level isolation.

How to eliminate wrong answers

Option A is wrong because using a service endpoint with a managed identity only secures network access at the subnet level and provides identity-based authentication, but it does not restrict the endpoint to a specific virtual network in the same way a private endpoint does, and it does not enable CMK for encryption at rest. Option B is wrong because disabling public network access alone does not restrict access to a specific virtual network; it only blocks all public traffic, and while enabling CMK via Azure Key Vault is correct for encryption, the network requirement is not met. Option C is wrong because configuring IP firewall rules only restricts access based on source IP addresses, not to a specific virtual network, and while CMK via Azure Key Vault is correct, the network isolation is insufficient for a VNet-specific restriction.

835
MCQmedium

A company is developing a conversational AI solution using Microsoft Copilot Studio. They want the copilot to answer questions based on a knowledge base of technical documents. Which data source integration should they use?

A.Azure AI Search
B.Azure Blob Storage
C.Azure SQL Database
D.Microsoft Lists
AnswerA

Azure AI Search can index documents and be used as a knowledge source.

Why this answer

The correct answer is B because Microsoft Copilot Studio can use Azure AI Search as a knowledge source for retrieval-augmented generation. A (Azure SQL Database) is not directly supported as a knowledge base for Copilot Studio. C (Azure Blob Storage) requires additional indexing.

D (Microsoft Lists) is for simpler data.

836
MCQeasy

A company is building an agent that needs to perform tasks like sending emails and updating a CRM system. The agent uses Azure OpenAI with function calling. The team defines functions for these tasks. When the agent is tested, it sometimes calls the wrong function or invents function names. What should the team do to improve the reliability of function calling?

A.Fine-tune the model on a dataset of correct function calls.
B.Reduce the number of functions to only the most common ones.
C.Set the temperature parameter to 0 for deterministic output.
D.Provide better function descriptions with examples of when to use each function.
AnswerD

Clear descriptions improve function selection.

Why this answer

Option D is correct because providing better function descriptions with examples directly improves the model's ability to select the appropriate function. Azure OpenAI's function calling relies on the semantic understanding of the function definitions; clear descriptions and usage examples reduce ambiguity, helping the model map user intent to the correct function signature without hallucinating names.

Exam trap

The trap here is that candidates often assume deterministic output (temperature=0) or reducing complexity (fewer functions) will fix reliability, when the real issue is semantic ambiguity in function definitions that the model cannot resolve without better descriptions.

How to eliminate wrong answers

Option A is wrong because fine-tuning on a dataset of correct function calls is unnecessary and inefficient; Azure OpenAI's base models already understand function calling patterns, and fine-tuning would require a large, curated dataset and could introduce overfitting or degrade general performance. Option B is wrong because reducing the number of functions limits the agent's capabilities and does not address the root cause of incorrect selection; the model may still invent names if descriptions are poor. Option C is wrong because setting temperature to 0 makes output deterministic but does not fix ambiguous or poorly defined function descriptions; the model will still confidently choose the wrong function if it misinterprets the intent.

837
MCQmedium

Refer to the exhibit. You are configuring an Azure AI Foundry agent for customer support. The agent uses Azure AI Search for retrieval and Azure OpenAI for generation. Users report that the agent provides correct answers but sometimes includes inappropriate language. What is the most likely cause?

A.The content safety blocklist is not applied to the chat output
B.The OpenAI deployment 'gpt-4o' does not support content filtering
C.The content safety threshold should be set to 'low' to block more content
D.The semantic configuration is not optimized for safety
AnswerA

Content safety configuration in the JSON is for input but may not be applied to the generated output; need to configure output filtering.

Why this answer

The content safety threshold is set to 'high', which is the most restrictive level. Option B is correct because 'high' threshold blocks more content, but the issue is inappropriate language, so the threshold might be too low if it allows offensive content; actually, 'high' is strict, so the issue is that the content safety is not applied to the generation output. Option A is wrong because semantic search is for relevance, not safety.

Option C is wrong because the deployment name is irrelevant. Option D is wrong because the system message does not guarantee safety filtering.

838
MCQhard

A bank uses Azure AI Document Intelligence to process loan applications. The solution must extract data from scanned PDFs and validate it against a database. The bank requires that all extracted data be encrypted at rest and in transit. Which security measure should you implement?

A.Enable customer-managed keys (CMK) with Azure Key Vault for the Document Intelligence resource
B.Use a system-assigned managed identity for the application
C.Configure a private endpoint for the Document Intelligence resource
D.Use Azure RBAC to restrict access to the Document Intelligence resource
AnswerA

CMK provides encryption at rest with customer-controlled keys.

Why this answer

Option C is correct because using a customer-managed key (CMK) with Azure Key Vault provides control over encryption keys and ensures data is encrypted at rest. Option A is wrong because Azure RBAC controls access but not encryption. Option B is wrong because private endpoint secures network traffic but does not encrypt data at rest.

Option D is wrong because managed identity is for authentication, not encryption.

839
Multi-Selectmedium

Which TWO features are available in Azure AI Language's extractive summarization?

Select 2 answers
A.Confidence scores for each extracted sentence.
B.Identified named entities from the document.
C.Sentiment scores for each extracted sentence.
D.Ranked list of sentences extracted from the document.
E.Generated abstractive summary.
AnswersA, D

Each sentence has a confidence score indicating its relevance.

Why this answer

Options A and E are correct because extractive summarization returns a ranked list of sentences and a confidence score for each sentence. Option B is wrong because it does not generate new sentences. Option C is wrong because it does not categorize sentences by sentiment.

Option D is wrong because it does not identify named entities; that's entity recognition.

840
Multi-Selectmedium

Which TWO actions can reduce the cost of using Azure Custom Vision for image classification? (Choose two.)

Select 2 answers
A.Include negative samples in the dataset.
B.Use the compact domain for faster training.
C.Reduce the number of training images.
D.Use a larger image size for higher accuracy.
E.Increase the number of training iterations.
AnswersB, C

Compact domain reduces training time and cost.

Why this answer

Option B is correct because using the compact domain in Azure Custom Vision reduces model complexity and training time, which directly lowers compute costs. Compact domains are optimized for edge deployment and require fewer resources, making them more cost-effective for image classification tasks.

Exam trap

The trap here is that candidates often confuse 'faster training' with 'reduced cost' but may overlook that compact domains specifically lower resource consumption, while options like reducing images or iterations seem intuitive but are not explicitly cost-reduction features in the exam context.

841
MCQhard

A PowerShell script (exhibit) attempts to update the endpoint for an Azure AI Language resource. What is the outcome of running this script?

A.The script creates a new resource
B.The script changes the API version
C.The script updates the endpoint successfully
D.The script fails because the property does not exist
AnswerD

The property path is invalid, causing an error.

Why this answer

The property path 'apiProperties.textAnalytics.documentEndpoint' is incorrect. The correct property for custom endpoint is not modifiable via this method; the script will fail because the property does not exist. The endpoint is set during creation and cannot be changed this way.

842
MCQhard

You are using Azure AI Language's conversational language understanding (CLU). The above JSON is a request to a CLU endpoint. What is the purpose of this request?

A.To predict the intent and entities from the user utterance
B.To query a knowledge base for answers
C.To deploy the CLU model to production
D.To train a new CLU model
AnswerA

The analysisInput contains the utterance for prediction.

Why this answer

The JSON request is sent to the Azure AI Language CLU endpoint with a 'query' field containing the user utterance. The 'kind' field is set to 'Conversation', which triggers the CLU runtime to analyze the utterance against the deployed model. The purpose is to return a prediction of the top intent and any extracted entities, which is the core function of a conversational language understanding endpoint.

Exam trap

The trap here is that candidates confuse the CLU prediction endpoint with the training or deployment endpoints, mistakenly thinking a request with a 'query' field is used for model management rather than runtime inference.

How to eliminate wrong answers

Option B is wrong because querying a knowledge base for answers is the purpose of Azure AI Language's custom question answering (QnA Maker) or Azure Cognitive Search, not CLU. Option C is wrong because deploying a CLU model is a separate operation performed via the Azure portal, REST API (e.g., PUT on the deployment resource), or SDK; this request is a prediction call, not a deployment action. Option D is wrong because training a new CLU model requires a training API call (e.g., POST to the /train endpoint with a training dataset), not a prediction request to the runtime endpoint.

843
MCQhard

Your Azure AI Search index contains millions of documents. Users report that search results are slow for complex queries. You need to improve query performance without reducing result quality. Which action should you take?

A.Reduce the maximum number of results returned per query
B.Increase the number of replicas
C.Remove all facet fields from the index
D.Disable complex query types such as fuzzy and regex
AnswerB

Adding replicas allows load balancing and faster query responses.

Why this answer

Increasing the number of replicas in Azure AI Search distributes query load across multiple copies of the index, enabling parallel processing of complex queries. This directly improves query throughput and latency without altering the index schema or reducing result quality, as replicas provide dedicated resources for query execution.

Exam trap

The trap here is that candidates confuse replicas (which improve query performance and availability) with partitions (which improve indexing speed and storage capacity), leading them to choose options that degrade functionality instead of scaling resources.

How to eliminate wrong answers

Option A is wrong because reducing the maximum number of results per query (e.g., via $top) only limits the response payload and does not address the underlying computational cost of complex queries; it can also degrade user experience by hiding relevant results. Option C is wrong because removing facet fields eliminates aggregation capabilities and does not improve query performance—facets are computed during indexing, not at query time, and their removal would reduce result quality by removing navigation aids. Option D is wrong because disabling complex query types (fuzzy, regex) restricts search functionality and may reduce result relevance; while these queries are resource-intensive, the correct approach is to scale out via replicas rather than sacrifice search capabilities.

844
MCQmedium

Refer to the exhibit. You have deployed a GPT-3.5 Turbo model in Azure OpenAI Service with the shown configuration. Users report that the model generates responses that are too repetitive. You need to reduce repetition. Which parameter should you modify?

A.Increase presencePenalty to 0.5
B.Increase frequencyPenalty to 0.5
C.Increase temperature to 1.0
D.Decrease topP to 0.5
AnswerA

Presence penalty reduces the likelihood of repeating any token that has appeared, reducing repetition.

Why this answer

Option D is correct because presence_penalty reduces the likelihood of repeating tokens that have already appeared. Option A is wrong because temperature controls creativity. Option B is wrong because topP controls nucleus sampling.

Option C is wrong because frequency_penalty is set to 0 and could be increased, but presence_penalty is more effective for repetition of topics. However, both can reduce repetition; but presence_penalty is often preferred for reducing repetition of content. Based on typical usage, presence_penalty is the better choice.

845
MCQeasy

You are implementing a chatbot using Microsoft Copilot Studio that helps employees find company policies. The chatbot must: - Use generative answers based on a SharePoint Online site. - Only respond with information from approved policy documents. - Include citations in responses. - Be accessible from Microsoft Teams. - Require no custom code. What should you do?

A.Use Power Automate to retrieve documents and feed them to Azure OpenAI. Build a custom connector for Teams.
B.In Copilot Studio, create a new copilot. Add the SharePoint site as a knowledge source. Enable generative answers with citations. Publish to Teams.
C.Build a bot using Azure Bot Service and QnA Maker. Train it with the policy documents. Deploy to Teams.
D.Create a custom GPT in Azure OpenAI Studio. Upload the policy documents. Deploy via Azure API Management and expose to Teams.
AnswerB

Simplest approach meeting all requirements.

Why this answer

Option B is correct because Microsoft Copilot Studio natively supports adding a SharePoint Online site as a knowledge source, enabling generative answers that retrieve and cite only approved policy documents. It requires no custom code, automatically includes citations in responses, and can be published directly to Microsoft Teams, fulfilling all stated requirements.

Exam trap

The trap here is that candidates may overcomplicate the solution by choosing Azure OpenAI or Azure Bot Service options, missing that Copilot Studio is the no-code, fully integrated tool designed specifically for this scenario with built-in SharePoint knowledge sources, citations, and Teams deployment.

How to eliminate wrong answers

Option A is wrong because it requires custom code (Power Automate flow, custom connector) and Azure OpenAI, which violates the 'no custom code' requirement and adds unnecessary complexity. Option C is wrong because QnA Maker is deprecated and does not support generative answers with citations from SharePoint; it also requires manual training and custom deployment to Teams. Option D is wrong because creating a custom GPT in Azure OpenAI Studio and deploying via Azure API Management involves custom code and infrastructure management, contradicting the 'no custom code' and 'accessible from Teams' requirements without additional integration.

846
MCQeasy

Your organization needs to analyze customer feedback from social media posts to determine the sentiment (positive, negative, neutral). The solution must process up to 10,000 posts per day and provide a confidence score for each sentiment. Which Azure AI service should you use?

A.Azure AI Speech Service
B.Azure AI Language Service
C.Azure AI Translator
D.Azure AI Language Understanding (LUIS)
AnswerB

Offers sentiment analysis with confidence scores.

Why this answer

Option A is correct because Azure AI Language provides sentiment analysis with confidence scores as a built-in feature. Option B is wrong because the Translator service translates text, it does not analyze sentiment. Option C is wrong because the Speech service processes audio, not text.

Option D is wrong because the Language Understanding service (LUIS) is for intent and entity extraction, not sentiment analysis.

847
MCQmedium

You are using Azure OpenAI Service to generate marketing copy. The marketing team reports that the generated content sometimes contains factual inaccuracies. You need to improve the factual accuracy of the generated content. What should you do?

A.Increase the max_tokens parameter
B.Include relevant context and facts in the prompt
C.Decrease the temperature parameter
D.Disable content filtering
AnswerB

Providing context helps the model generate more accurate responses.

Why this answer

Option B is correct because providing relevant context in the prompt gives the model factual information to base its response on. Option A is wrong because reducing temperature does not improve factual accuracy. Option C is wrong because increasing max_tokens does not improve accuracy.

Option D is wrong because disabling content filtering does not address inaccuracies.

848
Multi-Selecthard

Which THREE factors should you consider when choosing between Azure AI Document Intelligence prebuilt models and custom models for invoice processing?

Select 3 answers
A.Both model types can be deployed on-premises.
B.Prebuilt models require no training data.
C.Prebuilt models are always less accurate than custom models.
D.Custom models require a large set of labeled training invoices.
E.Custom models can handle non-standard invoice layouts.
AnswersB, D, E

Prebuilt models are ready to use immediately.

Why this answer

Option B is correct because Azure AI Document Intelligence prebuilt models are designed to extract common fields from standard invoice layouts without requiring any labeled training data. They are pretrained on a large corpus of documents, enabling immediate use for typical invoice structures.

Exam trap

The trap here is that candidates assume prebuilt models are always less accurate than custom models, but accuracy depends on the document's similarity to the training data; prebuilt models can outperform custom ones on standard layouts, especially when training data is limited.

849
MCQeasy

You need to extract key-value pairs from scanned forms as part of a knowledge mining solution. Which Azure AI service should you use?

A.Azure AI Vision
B.Azure AI Language
C.Azure AI Search
D.Azure AI Document Intelligence
AnswerD

Specialized for form extraction.

Why this answer

Azure AI Document Intelligence (formerly Form Recognizer) is designed to extract key-value pairs from forms. Option B is incorrect because Vision is for OCR, not structure. Option C is incorrect because Language is for text analysis.

Option D is incorrect because Search is for indexing.

850
MCQeasy

You need to provide a business analyst with access to create and manage Azure AI Language projects without granting them full subscription-level permissions. What role should you assign?

A.Cognitive Services Language Owner at the resource level
B.Reader at the resource group level
C.Contributor at the subscription level
D.Cognitive Services User at the resource level
AnswerA

Grants full management of Language projects within the resource.

Why this answer

The Cognitive Services Language Owner role at the resource level grants full permissions to create, read, update, and delete Azure AI Language projects and resources, including managing custom models and deployments, without granting any permissions outside that specific resource. This is the least-privilege role that satisfies the business analyst's need to manage Language projects while avoiding subscription-level access.

Exam trap

The trap here is that candidates often confuse the Cognitive Services User role (which only allows consumption of the service) with the Language Owner role, mistakenly thinking 'User' implies management capabilities, or they default to a broad Contributor role without considering resource-level scoping.

How to eliminate wrong answers

Option B is wrong because the Reader role at the resource group level provides read-only access, which does not allow creating or managing projects. Option C is wrong because Contributor at the subscription level grants full management access to all resources in the subscription, far exceeding the required scope and violating the least-privilege principle. Option D is wrong because Cognitive Services User at the resource level only allows using the service (e.g., calling APIs) but does not include permissions to create or manage projects, which require owner-level or contributor-level roles.

851
Multi-Selecteasy

You are deploying a chat application using Azure OpenAI. The application should only answer questions based on a specific set of internal documents. Which THREE features should you use?

Select 3 answers
A.Azure AI Search index with the internal documents
B.Grounding with your data in Azure OpenAI Studio
C.Content filters to block out-of-domain questions
D.System message to limit the assistant's scope
E.Fine-tuning the model on the internal documents
AnswersA, B, D

The index provides the data source for grounding.

Why this answer

Option A is correct because Azure AI Search indexes allow you to ingest internal documents and perform vector or hybrid search over them. When integrated with Azure OpenAI, the search results are used as grounding context for the model, ensuring responses are based solely on your data.

Exam trap

Microsoft often tests the distinction between content filtering (which handles safety) and domain restriction (which requires retrieval or prompt engineering), leading candidates to incorrectly select content filters for limiting question scope.

852
MCQhard

You are a machine learning engineer at a retail company. The company wants to build a product knowledge base by extracting information from product manuals, specifications sheets, and customer reviews. The data sources include PDFs, Word documents, and plain text files stored in Azure Blob Storage. The solution must: (1) extract product name, model number, price, and key features; (2) analyze customer reviews to extract sentiment and common issues; (3) enable natural language queries like 'Which products have the best reviews under $100?'; (4) handle documents in English and Spanish. You need to design a solution using Azure AI Search and Azure AI Services. Which approach meets all requirements with the least development effort?

A.Use Azure AI Document Intelligence custom model to extract product info from manuals/specs. Use a separate Azure AI Search pipeline for customer reviews with sentiment analysis. Enable semantic search.
B.Use a single Azure AI Search pipeline with a skillset that includes Document Layout skill, Text Translation skill (to English), Sentiment skill, and Key Phrase Extraction skill. Enable semantic search.
C.Use Azure AI Search with a blob indexer and a skillset that includes OCR skill (for scanned PDFs), Text Translation skill, Sentiment skill, and Entity Recognition skill. Enable semantic search.
D.Use Azure AI Document Intelligence to extract product info from all documents, then feed into Azure AI Search. Enable semantic search.
AnswerB

Single pipeline handles all document types, translates, extracts sentiment, and enables natural language queries.

Why this answer

Option D is correct because one skillset can handle all document types (using Document Layout skill), translate Spanish to English, extract sentiment from reviews, and extract key phrases for features. Document Intelligence is not needed for reviews. Option A misses sentiment analysis.

Option B uses two pipelines unnecessarily. Option C misses translation.

853
MCQeasy

You need to generate an image of a cat wearing a hat using Azure OpenAI. Which model should you use?

A.Codex
B.DALL-E
C.GPT-4
D.Whisper
AnswerB

DALL-E generates images from text descriptions.

Why this answer

DALL-E is the Azure OpenAI model specifically designed for generating images from natural language descriptions. It uses a diffusion-based architecture to create high-quality, original images based on text prompts, making it the correct choice for generating an image of a cat wearing a hat.

Exam trap

The trap here is that candidates may confuse GPT-4's multimodal capabilities (which can analyze images but not generate them) with DALL-E's generative image creation, leading them to incorrectly select GPT-4 for image generation tasks.

How to eliminate wrong answers

Option A is wrong because Codex is a model specialized for generating code from natural language, not for image generation. Option C is wrong because GPT-4 is a large language model focused on text generation and reasoning, lacking native image generation capabilities. Option D is wrong because Whisper is a speech-to-text model designed for audio transcription, not image generation.

854
MCQhard

You are an AI engineer at a global e-commerce company. The company uses Azure AI Language to analyze product reviews in English, Spanish, and French. The current solution calls the sentiment analysis API for each review individually, resulting in high latency and cost. You need to design a new solution that processes reviews in batches, reduces the number of API calls, and still supports multiple languages. The solution must also extract key phrases and detect the language automatically. You have the following options: Option A: Use the Azure AI Language synchronous API with the 'multi-language' parameter set to true. Send reviews one by one. Option B: Use the Azure AI Language asynchronous batch API. Combine all reviews into a single batch request, but only for one language at a time. Option C: Use the Azure AI Language asynchronous batch API. Send a single batch request with all reviews, setting the 'language' parameter to 'multi' to auto-detect language, and specify sentiment analysis, key phrase extraction, and language detection as tasks. Option D: Use the Azure AI Translator service to translate all reviews to English, then use the Azure AI Language batch API for English-only sentiment and key phrase extraction.

A.Option C
B.Option B
C.Option A
D.Option D
AnswerA

Asynchronous batch API with multi-language and multiple tasks reduces calls and supports all languages.

Why this answer

Option C is correct. The asynchronous batch API supports multiple tasks (sentiment, key phrases, language detection) in a single request, and setting language to 'multi' enables auto-detection. This reduces API calls and latency.

Option A is wrong because synchronous one-by-one calls increase latency/cost. Option B is wrong because it processes only one language per batch, requiring multiple batches. Option D is wrong because it adds translation cost and latency, and may lose nuances.

855
MCQeasy

A company is building a chatbot using Azure Cognitive Service for Language. They need to ensure that user utterances are correctly mapped to the appropriate intent in a custom question answering project. What should they configure?

A.Add synonyms and phrase list to the LUIS application.
B.Add alternative question phrases to the QnA pairs in the custom question answering project.
C.Define entities in the custom question answering project to capture key information.
D.Add synonyms and phrase list to the custom question answering project.
AnswerD

Synonyms and phrase lists help match varied user utterances to the correct QnA pair.

Why this answer

Option D is correct because adding synonyms and phrase lists to a custom question answering project directly improves the mapping of user utterances to intents by normalizing variations in phrasing. This configuration allows the project to recognize equivalent terms (e.g., 'cost' and 'price') as the same intent, ensuring accurate intent mapping without requiring exact matches.

Exam trap

The trap here is that candidates confuse the role of synonyms/phrase lists in custom question answering with the separate LUIS service, or mistakenly think alternative question phrases (Option B) are the primary mechanism for intent mapping, when in fact they only expand answer triggers, not intent classification.

How to eliminate wrong answers

Option A is wrong because LUIS (Language Understanding Intelligent Service) is a separate Azure service for intent and entity extraction, not used within custom question answering; the question specifies a custom question answering project, which does not use LUIS phrase lists. Option B is wrong because adding alternative question phrases to QnA pairs improves answer matching for specific questions but does not configure intent mapping across utterances—it only expands the set of questions that trigger a given answer. Option C is wrong because defining entities captures key information (e.g., dates, product names) from utterances but does not map utterances to intents; entities are for extracting data, not for intent classification.

856
Multi-Selectmedium

Which THREE factors should be considered when choosing between Azure Computer Vision and Azure Custom Vision? (Choose three.)

Select 3 answers
A.The need for custom model retraining over time.
B.Whether the solution runs on edge devices.
C.The amount of labeled training data available.
D.The geographic region of the Azure subscription.
E.Whether the detection objects are generic or domain-specific.
AnswersA, C, E

Custom Vision allows retraining.

Why this answer

Option A is correct because Azure Custom Vision is specifically designed for scenarios where you need to retrain a model over time with new labeled data, such as when the visual characteristics of objects change (e.g., new product packaging). Azure Computer Vision is a pre-trained API that cannot be retrained; it only supports fixed, generic models. Custom Vision allows iterative training with your own images, making it essential when model drift or evolving requirements demand periodic retraining.

Exam trap

Microsoft often tests the misconception that edge deployment is exclusive to Custom Vision, but in reality, both services support containerized edge deployment, so the true differentiator is the need for custom retraining and domain-specific detection.

857
MCQeasy

You need to deploy a generative AI model that can be used by multiple applications within your organization. The model must support real-time inference with low latency. Which Azure service should you use?

A.Azure AI Search
B.Azure OpenAI Service
C.Azure Machine Learning real-time endpoint
D.Azure Functions
AnswerB

Azure OpenAI Service provides managed endpoints with low-latency inference for generative AI models.

Why this answer

Option B is correct because Azure OpenAI Service provides managed endpoints with low latency for real-time inference. Option A is wrong because Azure Machine Learning real-time endpoints also work but require more management overhead; Azure OpenAI is more purpose-built for generative AI. Option C is wrong because Azure AI Search is for indexing and search, not inference.

Option D is wrong because Azure Functions is a compute service, not designed for hosting large language models.

858
Multi-Selectmedium

You need to design a computer vision solution that detects defects in manufactured parts on a conveyor belt. The solution must run in near real-time and adapt to new defect types without retraining from scratch. Which TWO approaches should you consider?

Select 2 answers
A.Use Azure AI Face to detect anomalies
B.Use Azure AI Custom Vision with object detection and retrain with new defect images
C.Implement transfer learning with a pre-trained model and fine-tune on defect images
D.Use Azure AI Video Indexer to analyze video feeds
E.Use pre-built Azure AI Vision Image Analysis to classify images
AnswersB, C

Custom Vision supports retraining with new images to learn new defects.

Why this answer

Custom Vision with object detection can be retrained with new defect images. Transfer learning allows quick adaptation to new defect types.

859
MCQmedium

Refer to the exhibit. A developer received this response from an Azure OpenAI chat completion call. The prompt was "What is the capital of France?". The finish_reason is "stop". What does this indicate?

A.The response was truncated due to content filtering.
B.The model completed the response naturally.
C.The model stopped generating before the response was complete.
D.The response reached the max_tokens limit.
AnswerB

Finish_reason 'stop' indicates normal completion.

Why this answer

The finish_reason 'stop' indicates that the model completed the response naturally, meaning it generated a complete answer to the prompt and reached a logical stopping point (e.g., the end of a sentence or the end of the generated text). This is the standard behavior for a successful completion where the model did not encounter any content filter, token limit, or other interruption.

Exam trap

Microsoft often tests the distinction between finish_reason values, and the trap here is that candidates confuse 'stop' with 'length' or assume any non-error finish_reason means truncation, when in fact 'stop' explicitly signals a natural and complete generation.

How to eliminate wrong answers

Option A is wrong because 'stop' specifically means the model finished generating on its own, not that content filtering truncated the response; content filtering would return a finish_reason of 'content_filter'. Option C is wrong because 'stop' indicates the model completed the response, not that it stopped prematurely; a premature stop would be indicated by a finish_reason of 'length' (if max_tokens hit) or 'null' (if interrupted). Option D is wrong because reaching the max_tokens limit would result in a finish_reason of 'length', not 'stop'.

860
MCQmedium

Your company uses Azure AI Vision to analyze images. You receive an alert that the number of 429 (Too Many Requests) errors has increased significantly. What is the most likely cause?

A.The endpoint URL is incorrect.
B.The API key has expired.
C.The service principal does not have the correct role assignment.
D.The application is exceeding the transactions-per-second (TPS) limit.
AnswerD

429 errors occur when the request rate exceeds the allowed TPS.

Why this answer

HTTP 429 (Too Many Requests) is a rate-limiting response from Azure AI Vision when the client exceeds the allowed transactions-per-second (TPS) for the chosen pricing tier. The alert indicates the application is sending requests faster than the service's capacity, triggering throttling to protect backend resources.

Exam trap

The trap here is confusing HTTP 429 with authentication or authorization errors (401/403), leading candidates to incorrectly select options about API keys or role assignments when the real issue is rate limiting.

How to eliminate wrong answers

Option A is wrong because an incorrect endpoint URL would produce a 404 Not Found or connection error, not a 429 rate-limit error. Option B is wrong because an expired API key results in a 401 Unauthorized or 403 Forbidden response, not a 429. Option C is wrong because an incorrect role assignment on the service principal would cause 403 Forbidden errors due to missing RBAC permissions, not a 429 throttling response.

861
MCQhard

A legal compliance team needs to automatically redact personally identifiable information (PII) from legal documents before sharing them with external auditors. The documents are stored in Azure Blob Storage. The solution must use Azure AI Language to detect PII and then redact the identified entities. The redaction must be performed on the original documents, and the redacted versions must be saved to a separate container. You need to design a serverless solution with minimal latency. What should you do?

A.Use Azure Data Factory to copy documents to a processing location and call an Azure Function.
B.Use Azure Batch Service to process documents in parallel and redact PII.
C.Create an Azure Function triggered by Blob Storage events to detect PII, redact, and save to a separate container.
D.Use Azure Logic Apps with a Blob trigger to call the PII detection API and write redacted documents.
AnswerC

Azure Functions provide low-latency, serverless event-driven processing.

Why this answer

Azure Functions can be triggered by Blob Storage events, process each document using the PII detection API, redact the entities, and save the redacted version to a separate container. This is serverless and efficient. Option C is correct.

Option A is wrong because Logic Apps has higher latency due to connectors. Option B is wrong because Data Factory is for data movement. Option D is wrong because Batch Service is for large-scale batch processing, not real-time.

862
Multi-Selecteasy

Which TWO Azure AI services can be used to extract text from images as part of a knowledge mining pipeline?

Select 2 answers
A.Azure AI Language
B.Azure AI Document Intelligence
C.Azure AI Computer Vision
D.Azure AI Video Indexer
E.Azure AI Custom Vision
AnswersB, C

Includes OCR and layout extraction.

Why this answer

Azure AI Document Intelligence (formerly Form Recognizer) is correct because it is specifically designed to extract text, tables, and key-value pairs from scanned documents and images using optical character recognition (OCR) and deep learning models. It is a core service for knowledge mining pipelines that require structured data extraction from unstructured documents.

Exam trap

The trap here is that candidates often confuse Azure AI Computer Vision's OCR capabilities with Azure AI Document Intelligence, but Document Intelligence is the dedicated service for structured document extraction in knowledge mining, while Computer Vision provides general-purpose image analysis and OCR without the same level of document-specific parsing.

863
MCQmedium

You are the AI engineer at a global e-commerce company that allows users to upload product images and descriptions. You use Azure Content Moderator to automatically moderate images for adult and racy content, and text for profanity and personal data. Recently, you noticed that some product descriptions containing profanity in French are not being flagged. Your Content Moderator text moderation API call includes the language parameter set to 'eng'. The profanity list appears to be English-only. You have a requirement to support French and Spanish in addition to English. You also need to ensure that false positives for legitimate product descriptions are minimized. You cannot use a custom term list because the profanity terms are dynamic. What should you do?

A.Disable the language parameter so that the API defaults to all languages.
B.Create separate API calls for each language and specify the language code in the request.
C.Set the language parameter to 'auto-detect' in the text moderation API request.
D.Add French and Spanish profanity terms to a custom term list and use the list in the API call.
AnswerC

Auto-detect enables Content Moderator to identify the language and apply the correct profanity detection model, supporting multiple languages dynamically.

Why this answer

Option C is correct because setting the language parameter to 'auto-detect' allows the Azure Content Moderator text moderation API to automatically identify the language of the input text and apply the corresponding built-in profanity list (including French and Spanish). This meets the requirement to support multiple languages without using a custom term list, and it minimizes false positives by using the appropriate language-specific moderation model rather than a generic English-only list.

Exam trap

The trap here is that candidates may think disabling the language parameter or creating separate calls will enable multi-language support, but they overlook that the API's default behavior is English-only unless 'auto-detect' is explicitly specified, and that custom term lists are not allowed per the scenario's constraints.

How to eliminate wrong answers

Option A is wrong because disabling the language parameter does not cause the API to default to all languages; instead, it defaults to English-only moderation, which would still miss French and Spanish profanity. Option B is wrong because creating separate API calls for each language is inefficient and does not solve the core issue—the API still uses the English-only profanity list unless the language parameter is set to 'auto-detect' or a supported language code. Option D is wrong because the requirement explicitly states you cannot use a custom term list due to dynamic profanity terms, and adding French and Spanish terms to a custom list would violate that constraint and introduce maintenance overhead.

864
MCQeasy

A company wants to use Azure AI Language to automatically summarize large documents. The summarization must extract the most important sentences from each document. Which feature should they use?

A.Extractive summarization
B.Abstractive summarization
C.Key phrase extraction
D.Entity recognition
AnswerA

Extracts important sentences.

Why this answer

Option B is correct because extractive summarization selects key sentences from the document. Option A is wrong because abstractive summarization generates new sentences, not extracts. Option C is wrong because key phrase extraction extracts phrases, not sentences.

Option D is wrong because entity recognition extracts named entities.

865
Matchingmedium

Match each Azure Cognitive Services endpoint to its purpose.

Drag a concept onto its matching description — or click a concept then click the description.

Concepts
Matches

Analyze sentiment of text

Generate description of an image

Query a knowledge base

Detect faces in an image

Translate text between languages

Why these pairings

These are common REST API endpoints for Azure Cognitive Services.

866
MCQhard

You are the Azure AI engineer for a large e-commerce company. The company uses Azure Computer Vision to automatically tag product images uploaded by sellers. The system has been running smoothly for months. However, after a recent update to the Computer Vision API, you notice that certain images of clothing items are being tagged with incorrect labels, such as 'shoe' for a shirt. The images are clear and well-lit. You have confirmed that the image format (JPEG) is supported and the size is within limits. The issue occurs consistently for clothing items with similar colors. Other product categories work fine. You suspect the issue is related to the API version. What should you do first?

A.Increase the image size limit.
B.Check the API version used in the application code and compare with the latest version.
C.Switch to a custom model trained on clothing items.
D.Reduce the confidence threshold to 50% to see if more tags appear.
AnswerB

API updates may change behavior; rolling back or updating code may fix.

Why this answer

Option B is correct because the issue began after a Computer Vision API update, and the problem is specific to certain clothing images with similar colors, indicating a potential regression or behavioral change in the API version. Checking the API version used in the application code against the latest version is the first logical troubleshooting step to identify if a breaking change or bug was introduced. This aligns with Azure AI best practices: always verify API version compatibility before modifying thresholds or retraining models.

Exam trap

The trap here is that candidates may jump to retraining a custom model (Option C) or adjusting confidence thresholds (Option D) without first verifying the API version, which is the simplest and most cost-effective diagnostic step in Azure AI troubleshooting.

How to eliminate wrong answers

Option A is wrong because increasing the image size limit does not address incorrect labeling; the images are already within supported size limits and the issue is with tag accuracy, not file size. Option C is wrong because switching to a custom model is a significant investment and should only be considered after verifying that the pre-built API version is not the root cause; the problem may be a temporary API regression that a version rollback could fix. Option D is wrong because reducing the confidence threshold to 50% would increase the number of tags but not correct the mislabeling of a shirt as 'shoe'; it would likely introduce more false positives without fixing the core issue.

867
MCQeasy

Your team has built a knowledge mining pipeline using Azure AI Search and Document Intelligence. After ingestion, you notice that some documents are not appearing in search results. What is the most likely cause?

A.The indexer encountered errors and marked the documents as failed
B.The index does not have a semantic configuration
C.The search service has insufficient replicas
D.The search service is throttled due to high query volume
AnswerA

Indexer errors prevent documents from being indexed.

Why this answer

Option B is correct because if the indexer fails, documents are not indexed. Option A is wrong because throttling would affect all documents, not a subset. Option C is wrong because the search service would report failures for other reasons.

Option D is wrong because missing semantic configuration affects ranking, not indexing.

868
MCQhard

You are deploying a Custom Vision object detection model to an Azure Container Instance for real-time inference. The model must respond within 500 ms. The default container runs on CPU. What should you do to meet the latency requirement?

A.Increase the number of CPU cores in the container instance.
B.Export the model as a Dockerfile with GPU support and deploy to a GPU-enabled ACI.
C.Deploy the model to Azure Functions with a Premium plan.
D.Use the Cognitive Services Computer Vision container instead.
AnswerB

GPU acceleration is key for low-latency object detection.

Why this answer

Option D is correct because using a GPU-optimized container and a GPU-enabled Azure Container Instance will significantly reduce inference time. Increasing the number of CPU cores (A) helps but may not meet 500 ms for object detection. Using Azure Functions (B) adds cold start latency.

Using Cognitive Services container (C) is not correct for Custom Vision export.

869
Multi-Selectmedium

Which TWO Azure AI services can perform optical character recognition (OCR)?

Select 2 answers
A.Custom Vision
B.Azure AI Document Intelligence
C.Face API
D.Video Indexer
E.Read API
AnswersB, E

Document Intelligence includes OCR for document processing.

Why this answer

Options A and D are correct. The Read API provides OCR for images and documents. Azure AI Document Intelligence includes OCR for documents.

Option B is wrong because Face API does not do OCR. Option C is wrong because Custom Vision is for image classification. Option E is wrong because Video Indexer extracts text from video, but the question is about OCR in general; however, the Read API and Document Intelligence are the primary OCR services.

870
Multi-Selectmedium

Which TWO actions should you take to optimize the performance of an Azure AI Search solution that indexes large volumes of data?

Select 2 answers
A.Use the appropriate search tier (S1, S2, etc.) based on document size
B.Use the free tier for production workloads
C.Increase the replica count for better indexing throughput
D.Batch documents in groups of up to 1000 per index operation
E.Disable scoring profiles to speed up indexing
AnswersA, D

Higher tiers have better indexing capacity.

Why this answer

Option A is correct because choosing the appropriate search tier (S1, S2, etc.) ensures that the service has sufficient resources (CPU, memory, disk I/O) to handle the indexing load and document size. Higher tiers provide better indexing throughput and storage capacity, which is critical for large volumes of data. Using an undersized tier can lead to throttling, timeouts, or failed indexing operations.

Exam trap

The trap here is confusing replicas (which scale query performance) with partitions (which scale indexing throughput), leading candidates to incorrectly select option C as a way to improve indexing speed.

871
Multi-Selecteasy

Which TWO Azure services can be used together with Azure OpenAI to implement a Retrieval-Augmented Generation (RAG) solution?

Select 2 answers
A.Azure Blob Storage
B.Azure Cosmos DB
C.Azure AI Search
D.Azure SQL Database
E.Azure Functions
AnswersB, C

Cosmos DB supports vector indexing.

Why this answer

Azure Cosmos DB is correct because it can serve as a vector database for storing and querying embeddings generated by Azure OpenAI. When combined with Azure AI Search, which provides hybrid search (vector + keyword) and indexing capabilities, it enables a complete RAG pipeline where relevant documents are retrieved and passed as context to the OpenAI model for grounded generation.

Exam trap

The trap here is that candidates often assume any storage service (like Blob Storage or SQL Database) can directly support RAG, but they overlook the need for native vector search or hybrid retrieval capabilities that only Azure AI Search and Cosmos DB (with vector support) provide.

872
MCQhard

You are building a generative AI application that uses Azure OpenAI Service. The application must handle sensitive user data and comply with GDPR. You need to ensure that the service does not store or log user prompts and completions. What should you configure?

A.Disable data logging in the Azure OpenAI Service resource
B.Deploy the resource in a GDPR-compliant region
C.Configure content filtering to block sensitive data
D.Use a custom endpoint that anonymizes data
AnswerA

Disabling data logging ensures prompts and completions are not stored.

Why this answer

Option C is correct because disabling data logging in the Azure OpenAI Service resource ensures that prompts and completions are not stored. Option A is wrong because the region does not affect logging. Option B is wrong because content filtering does not prevent logging.

Option D is wrong because the resource endpoint does not affect logging.

873
MCQmedium

You are building a customer support chatbot using Azure OpenAI Service. The chatbot must only respond based on the company's product documentation and should not generate answers outside that scope. Which approach should you use?

A.Implement content filters to block responses not found in the documentation.
B.Fine-tune a GPT-4 model on the product documentation.
C.Use Azure OpenAI On Your Data with a search index built from the documentation.
D.Use prompt engineering with a system message instructing the model to only answer from the documentation.
AnswerC

This approach grounds the model on the indexed documents, ensuring responses are based on the documentation.

Why this answer

Option C is correct because Azure OpenAI On Your Data with a search index ensures the model only generates responses grounded in the provided documents. Option A is wrong because fine-tuning alone does not prevent the model from generating ungrounded content. Option B is wrong because prompt engineering may still result in hallucination.

Option D is wrong because content filters block harmful content but do not enforce domain grounding.

874
MCQhard

Your company uses Azure AI Language to process legal documents. You have built a custom NER model to extract parties, dates, and obligations. The model performs well on English documents but now you need to support French and German documents. You have no labeled data in those languages. You want to use the existing English model as a starting point. The solution must be cost-effective and avoid manual labeling as much as possible. You also need to ensure that the model can be retrained quickly when new document types are added. Which approach should you take?

A.Use the prebuilt NER model for French and German to extract entities like dates and organizations, and then map them to your custom entity types.
B.Enable the multilingual option in the custom NER project and retrain the model using the existing English labeled data.
C.Use the English custom NER model as-is and translate the French and German documents to English before processing.
D.Build separate custom NER models for French and German by translating the English labeled data using Azure AI Translator and training new models.
AnswerB

The multilingual option allows the model to learn from English data and predict entities in French and German without additional labeled data, meeting cost-effectiveness and avoiding manual labeling.

Why this answer

Option C is correct because the multilingual option in custom NER allows training on English data and generalizing to other languages without additional labeled data. Option A is wrong because building separate models requires labeling in each language. Option B is wrong because translation of training data may not capture legal nuances.

Option D is wrong because using the prebuilt NER would not extract the custom entities (parties, dates, obligations in the specific format needed).

875
MCQeasy

Refer to the exhibit. You are creating an Azure Cognitive Services account using an ARM template snippet. What type of account is being created?

A.Azure AI Language
B.Azure AI Computer Vision
C.Azure AI Services multi-service account
D.Azure OpenAI Service
AnswerC

The kind 'CognitiveServices' creates a multi-service account.

Why this answer

Option C is correct. The 'kind' is 'CognitiveServices', which is a multi-service account that includes Computer Vision, Face, etc. If it were only Computer Vision, 'kind' would be 'ComputerVision'. 'Language' would be 'TextAnalytics'. 'OpenAI' would be 'OpenAI'.

876
MCQmedium

A financial services company uses Azure AI Document Intelligence to process loan applications. The solution extracts data from uploaded PDFs and stores it in Azure Cosmos DB. Recently, extraction errors increased due to poor-quality scans. The company needs to improve accuracy without manual intervention. The solution must also handle varying document layouts. You need to recommend a plan. What should you do?

A.Increase the throughput (TPS) of the Document Intelligence resource.
B.Enable OCR enhancement in Document Intelligence.
C.Switch to Azure AI Language for entity extraction.
D.Train a custom document extraction model using labeled samples of poor-quality scans and different layouts.
AnswerD

Custom models learn from specific examples, improving accuracy on varied layouts and quality.

Why this answer

Option D is correct because training a custom document extraction model using labeled samples of poor-quality scans and varying layouts directly addresses the root cause of extraction errors. Azure AI Document Intelligence's custom model capability allows you to train on specific document types and quality issues, improving accuracy without manual intervention. This approach adapts to the company's need to handle diverse layouts and degraded image quality, which prebuilt models cannot reliably manage.

Exam trap

The trap here is that candidates may confuse throughput scaling (Option A) or generic OCR enhancement (Option B) with actual model improvement, overlooking that only custom training can adapt to specific data quality issues and layout variations.

How to eliminate wrong answers

Option A is wrong because increasing throughput (TPS) only affects the rate of processing requests, not the accuracy of extraction from poor-quality scans; it does not improve model performance. Option B is wrong because OCR enhancement in Document Intelligence is a built-in feature that applies to all documents, but it cannot be specifically tuned to handle poor-quality scans or varying layouts; it does not replace the need for a custom model trained on the company's specific data. Option C is wrong because Azure AI Language is designed for text analytics and entity extraction from clean text, not for processing scanned PDFs with layout variations and poor image quality; it lacks the document understanding capabilities of Document Intelligence.

877
MCQhard

A company uses Azure Document Intelligence to extract data from invoices. They deploy the model to a container for on-premises processing. After deployment, they notice that the container consumes more memory than expected. What should they do to optimize memory usage?

A.Set the 'Memory' environment variable to a lower value in the container configuration
B.Use the 'Read' model instead of the 'Layout' model
C.Use the cloud API instead of the container
D.Reduce the batch size in the client application
AnswerA

The container's memory usage can be controlled via the 'Memory' setting.

Why this answer

Option A is correct because Azure Document Intelligence containers expose a 'Memory' environment variable that allows you to limit the container's memory allocation. By setting this variable to a lower value, you constrain the container's memory usage, which directly addresses the issue of higher-than-expected consumption. This is the recommended approach for optimizing memory in containerized deployments.

Exam trap

The trap here is that candidates may think memory optimization is achieved by changing the model or client-side settings, when in fact container memory is controlled by environment variables or runtime resource limits specific to the container orchestrator.

How to eliminate wrong answers

Option B is wrong because switching from the 'Layout' model to the 'Read' model changes the extraction capabilities (e.g., losing table and structure extraction) but does not directly control or reduce the container's memory allocation; memory usage is governed by container runtime settings, not model selection. Option C is wrong because moving to the cloud API would change the deployment model entirely, not optimize memory usage of the existing container; the question specifically asks about optimizing the container deployment. Option D is wrong because reducing the batch size in the client application affects throughput and request size, not the memory footprint of the container itself; the container's memory consumption is primarily determined by its internal processes and allocated resources, not client-side batching.

878
MCQmedium

You need to build a chatbot that answers questions based on your company's internal knowledge base. The knowledge base consists of Word documents and PDFs. Which service should you use to create a conversational interface that retrieves answers from these documents?

A.Azure AI Search with Azure AI Bot Service
B.Azure AI Language Service - Custom Question Answering
C.Azure AI Computer Vision
D.Azure AI Document Intelligence
AnswerA

Index documents with Search and use Bot Service for Q&A.

Why this answer

Option D is correct because Azure AI Search can index documents and be used as a data source for a question-answering system, and Azure AI Bot Service can host the chatbot. Option A is wrong because Language Service's QnA Maker (now Custom Question Answering) can use a knowledge base but not directly index documents; it requires a search index for large datasets. Option B is wrong because Document Intelligence extracts data but does not answer questions.

Option C is wrong because Computer Vision does not handle text queries.

879
MCQeasy

You deploy a custom vision model using Azure AI Custom Vision. After deployment, you notice the model has high accuracy on training data but low accuracy on new images. What is the most likely cause?

A.The training time was too short
B.The training dataset has too few images
C.The wrong domain was selected during training
D.The model is overfitted to the training data
AnswerD

Overfitting leads to high training accuracy but poor generalization.

Why this answer

High accuracy on training data but low accuracy on new images is the classic symptom of overfitting, where the model has memorized the training examples (including noise and irrelevant patterns) rather than learning generalizable features. In Azure AI Custom Vision, this typically occurs when the training dataset is too small, too homogeneous, or lacks sufficient variation, causing the model to fail on unseen data.

Exam trap

The trap here is that candidates confuse 'too few images' (a contributing factor) with the direct diagnosis of 'overfitting,' but the question asks for the most likely cause of the described symptom, which is the overfitting itself, not its root cause.

How to eliminate wrong answers

Option A is wrong because training time in Custom Vision is automatically managed by the service; extending it does not directly cause overfitting—the model stops when convergence is reached. Option B is wrong because having too few images can contribute to overfitting, but the question asks for the 'most likely cause' given the symptom, and overfitting is the direct description of the behavior, not the root cause of small dataset size. Option C is wrong because selecting the wrong domain (e.g., 'General' vs. 'Food' or 'Landmarks') affects feature extraction and may reduce accuracy overall, but it does not specifically produce the pattern of high training accuracy and low test accuracy—that pattern is the hallmark of overfitting.

880
Multi-Selecteasy

Which TWO statements about Azure OpenAI Service content filters are true?

Select 2 answers
A.They can be configured with severity levels (low, medium, high)
B.They only filter the output of the model
C.They cannot be customized for specific use cases
D.They are bypassed when using PTU deployments
E.They include categories such as hate, sexual, violence, and self-harm
AnswersA, E

Severity levels allow granular control over filtering.

Why this answer

Option A is correct because Azure OpenAI Service content filters can be configured with severity levels (low, medium, high) to control the strictness of filtering for each content category. This allows administrators to fine-tune the filter sensitivity based on their application's risk tolerance and compliance requirements.

Exam trap

The trap here is that candidates often assume content filters only apply to model outputs (Option B) or that PTU deployments offer a way to bypass safety controls (Option D), but Azure enforces filters uniformly across all deployment types.

881
MCQmedium

You are building a multilingual chatbot using Azure AI Language. For a given user utterance, you need to first detect the language, then route to the appropriate language-specific intent model. Which combination of Azure AI Language features should you use?

A.Language Detection and Conversational Language Understanding
B.Key Phrase Extraction and Conversational Language Understanding
C.Translator and Conversational Language Understanding
D.Language Detection and Custom Text Classification
AnswerA

Language Detection identifies the language, then CLU processes intents in that language.

Why this answer

Language Detection identifies the language, then CLU (Conversational Language Understanding) processes intents. Custom Text Classification is for document-level classification, not conversational intents. Translator translates text but does not detect language.

Key Phrase Extraction is not needed.

882
MCQhard

Your company uses Azure Bot Service with QnA Maker to answer HR questions. You need to migrate to Microsoft Copilot Studio to leverage generative AI capabilities. You have an existing QnA Maker knowledge base. What is the recommended migration path?

A.In Copilot Studio, use the 'Import from QnA Maker' option to directly bring in the knowledge base.
B.Create a new Copilot and manually recreate each Q&A pair.
C.Use Azure PowerShell to export the knowledge base and then upload to Copilot Studio.
D.Export the QnA Maker KB to a TSV file, then import into Azure AI Language custom question answering.
AnswerA

Copilot Studio provides built-in import from QnA Maker.

Why this answer

Option C is correct because Copilot Studio supports importing QnA Maker knowledge bases directly. Option A is wrong because starting from scratch is unnecessary. Option B is wrong because Azure AI Language custom question answering is a different service; Copilot Studio has native import.

Option D is wrong because manual entry is inefficient.

883
MCQeasy

You are planning an Azure AI solution that uses Azure AI Document Intelligence to extract data from invoices. The solution must handle large volumes of documents with varying layouts. Which pricing tier should you choose to ensure scalability and high throughput?

A.Premium (P0)
B.Standard (S0)
C.Free (F0)
D.Basic (B0)
AnswerB

Standard tier provides scalable throughput for production workloads.

Why this answer

The Standard (S0) tier is correct because Azure AI Document Intelligence requires the S0 tier for production workloads that demand high throughput and scalability across varying document layouts. The S0 tier supports unlimited transactions per second (TPS) with auto-scaling, whereas lower tiers impose strict rate limits or lack the capacity to handle large volumes of invoices with diverse formats.

Exam trap

The trap here is that candidates may confuse Azure AI Document Intelligence's tier structure with other Azure AI services (like Cognitive Services) that offer a 'Premium' tier, or assume a 'Basic' tier exists for cost savings, when in fact Document Intelligence only has Free and Standard tiers.

How to eliminate wrong answers

Option A is wrong because Azure AI Document Intelligence does not offer a 'Premium (P0)' tier; the service uses Free (F0) and Standard (S0) tiers only. Option C is wrong because the Free (F0) tier is limited to 20 transactions per minute and is intended for evaluation and development, not for production-scale invoice processing with high throughput. Option D is wrong because Azure AI Document Intelligence does not have a 'Basic (B0)' tier; the only paid tier is Standard (S0), which provides the necessary scalability and throughput for large volumes.

884
MCQmedium

You are designing a knowledge mining solution for a manufacturing company that needs to extract information from equipment maintenance manuals. The manuals are in multiple languages (English, French, German). You need to ensure that the extracted content is searchable in English only. Which approach should you use?

A.Use the Entity Recognition skill to extract entities and then index entities only.
B.Use the Language Detection skill to identify language and then index all content as-is.
C.Use the Text Translation skill to translate all content to English during indexing.
D.Use the Key Phrase Extraction skill to extract key phrases and then index them.
AnswerC

Text Translation skill translates documents to a target language, enabling search in English only.

Why this answer

Option C is correct because you can use the Text Translation skill to translate content to English during indexing, and then index only the translated text. Option A would not translate. Option B only detects language.

Option D uses two skills unnecessarily.

885
MCQhard

You have a custom Named Entity Recognition (NER) model trained using Azure AI Language. The model is performing poorly on new data. You need to improve its accuracy. Which action should you take first?

A.Increase the training epochs.
B.Retrain the model using the same training data.
C.Review the test set results and add more labeled examples for entities with low precision/recall.
D.Reduce the number of entity types in the model.
AnswerC

Adding targeted training data helps the model learn patterns it missed.

Why this answer

Option B is correct because the first step to improve model accuracy is to review the test set results and identify errors, then add more labeled examples for those failing patterns. Option A is wrong because reducing the number of entities might lose important distinctions. Option C is wrong because retraining with the same data will not fix issues.

Option D is wrong because increasing epochs may lead to overfitting without better data.

886
MCQhard

You are designing a solution that must extract specific entities from customer emails, such as product names, order numbers, and dates. The solution must be able to learn from a small set of labeled examples and improve over time. Which Azure AI service should you use?

A.QnA Maker
B.Pre-built Entity Extraction
C.Text Analytics for health
D.Custom Entity Extraction
AnswerD

Trainable with small labeled datasets to extract custom entities.

Why this answer

Option C is correct because Custom Entity Extraction in Azure AI Language allows you to train a model with a small set of labeled data to extract custom entities. Option A is wrong because Pre-built Entity Extraction only works with predefined entity types. Option B is wrong because Text Analytics for health is specialized for medical data.

Option D is wrong because the QnA Maker is for question answering, not entity extraction.

887
Multi-Selecthard

Which THREE components are essential when building a custom skill for Azure AI Search?

Select 3 answers
A.A Web API endpoint that processes documents
B.Field mappings to pass data between the skill and the indexer
C.A machine learning model trained in Azure Machine Learning
D.An Azure Function to trigger the skill on a schedule
E.Input and output definitions in JSON format
AnswersA, B, E

Custom skills are implemented as web APIs.

Why this answer

A, B, and D are correct. A: The Web API is the endpoint. B: JSON input/output format is required.

D: Field mappings define inputs and outputs in the skillset. C is not required; E is optional.

888
Multi-Selecteasy

Which TWO Azure services can be used to perform optical character recognition (OCR) on documents? (Select two.)

Select 2 answers
A.Azure AI Metrics Advisor
B.Azure AI Language
C.Azure AI Document Intelligence
D.Azure AI Personalizer
E.Azure AI Vision
AnswersC, E

Extracts text and structure from documents.

Why this answer

Options A and C are correct. Azure AI Vision includes the Read API for OCR. Azure AI Document Intelligence also performs OCR as part of its document analysis.

Option B is wrong because Azure AI Language is for text analytics, not OCR. Option D is wrong because Azure AI Metrics Advisor is for anomaly detection. Option E is wrong because Azure AI Personalizer is for reinforcement learning.

889
Multi-Selecthard

Which THREE factors are critical to consider when designing a custom vision solution for a manufacturing quality inspection system?

Select 3 answers
A.Imbalance between defective and non-defective product samples.
B.Variation in lighting conditions across different inspection stations.
C.Inference latency requirements for real-time decisions.
D.The need for optical character recognition (OCR) of product serial numbers.
E.Multilingual support for labeling.
AnswersA, B, C

Class imbalance leads to biased models.

Why this answer

Option A is correct because class imbalance is a critical factor in custom vision solutions for manufacturing quality inspection. If defective samples are rare compared to non-defective ones, the model may become biased toward predicting the majority class, leading to poor recall for defects. Azure Custom Vision allows adjusting the probability threshold and using techniques like oversampling or weighted loss to mitigate this, but the imbalance must be accounted for during dataset preparation.

Exam trap

The trap here is that candidates may confuse peripheral requirements (like OCR or multilingual labels) with core design factors that directly impact model accuracy, latency, and robustness in a production vision system.

890
Multi-Selectmedium

Which TWO actions should you perform to ensure that an Azure AI Search indexer can successfully enrich documents using a custom skill that calls an external API?

Select 2 answers
A.Enable CORS on the Azure Function app to allow cross-origin requests
B.Configure a retry policy in the skillset definition for the custom skill
C.Provide a managed identity for the search service to access the Azure Function
D.Set the indexer's execution timeout to unlimited
E.Add the external API endpoint to the indexer's allowed domains list
AnswersB, C

Retry policy handles transient errors when calling the custom skill.

Why this answer

Options A and D are correct. A: Providing managed identity allows the indexer to authenticate to the Azure Function hosting the custom skill. D: Configuring a retry policy ensures transient failures are retried.

Option B is wrong because CORS is not needed for indexer-to-function calls (they are server-to-server). Option C is wrong because the indexer does not call the API directly; it calls the custom skill endpoint. Option E is wrong because the indexer already has its own timeout settings.

891
MCQmedium

You deploy an agent using the ARM template shown. Users report that the agent cannot answer questions about uploaded documents. What is the most likely cause?

A.The function tool is missing required parameters
B.The model specified is not supported for file operations
C.The file_search tool is disabled in the agent configuration
D.The code_interpreter tool is enabled, which conflicts with file_search
AnswerC

Disabled file_search prevents document search.

Why this answer

Option B is correct because 'file_search' is disabled in the template, so the agent cannot search uploaded documents. Option A is incorrect because code_interpreter is enabled. Option C is incorrect because the function is correctly defined.

Option D is incorrect because the model is valid.

892
MCQeasy

A developer is building a multilingual translation solution using Azure AI Translator. The solution must translate text to French, German, and Spanish. Which parameter should the developer set to specify the target language?

A.language
B.to
C.targetLanguage
D.from
AnswerB

'to' specifies target language.

Why this answer

Option B is correct because the 'to' parameter specifies the target language. Option A is wrong because 'from' specifies source language. Option C is wrong because 'language' is not a REST API parameter.

Option D is wrong because 'targetLanguage' is not the correct parameter name.

893
MCQhard

You are designing a multilingual chatbot using Azure AI Language. The chatbot must support English, Spanish, and French. You need to minimize development effort and ensure consistent intent recognition across languages. What should you do?

A.Use the Azure AI Translator to translate all utterances to English before sending to a single English-only CLU project.
B.Create a separate CLU project for each language and combine them with a routing mechanism.
C.Use a single CLU multilingual project that supports all three languages.
D.Use the Language Understanding (LUIS) service with a single app that includes language-specific utterances.
AnswerC

Multilingual project leverages cross-lingual transfer and reduces effort.

Why this answer

Option B is correct because Azure AI Language's multilingual project allows you to train a single model that works across multiple languages, leveraging cross-lingual transfer. Option A is wrong because creating separate projects for each language increases maintenance effort. Option C is wrong because the Translate API adds complexity and latency.

Option D is wrong because LUIS does not support multilingual projects as effectively as CLU.

894
MCQhard

You are deploying an Azure AI solution that uses Azure OpenAI Service. The solution must be deployed in a way that minimizes latency for users in Asia. However, the company's data residency policy requires data to stay in the United States. What should you do?

A.Use Azure CDN to cache the model responses in Asia.
B.Deploy the Azure OpenAI Service in an Asian region and use Azure Front Door to route traffic.
C.Deploy the service in multiple regions globally and use Traffic Manager for routing.
D.Deploy the service in a US region and use Azure Front Door with caching to reduce latency.
AnswerD

Front Door provides low-latency access while keeping data in US.

Why this answer

Option D is correct because it satisfies both requirements: data residency (deploying in a US region keeps data within the United States) and latency reduction for Asian users. Azure Front Door with caching stores frequently accessed model responses at edge locations closer to users in Asia, minimizing round-trip time without moving the origin data.

Exam trap

The trap here is that candidates assume caching (Option A) or global deployment (Option C) can solve latency without considering data residency, or they mistakenly think deploying in Asia (Option B) is acceptable despite the policy constraint.

How to eliminate wrong answers

Option A is wrong because Azure CDN caches static content, but Azure OpenAI Service responses are dynamic and often non-cacheable (e.g., unique prompts or streaming outputs); caching would not reduce latency for real-time inference. Option B is wrong because deploying in an Asian region violates the data residency policy requiring data to stay in the United States. Option C is wrong because deploying in multiple regions globally would require data replication outside the US, breaking the data residency constraint; Traffic Manager routes traffic but does not cache responses, so latency from a US region would remain high for Asian users.

895
MCQmedium

You are building a custom entity extraction solution using Azure AI Language. You have a small dataset (50 documents) with annotated entities. You need to train a model that can extract similar entities from new documents. What is the best approach?

A.Create a custom NER project in Azure AI Language and train it with your annotated data.
B.Use the prebuilt entity recognition API to extract entities.
C.Use the Conversational PII entity extraction feature.
D.Use the built-in entity extraction skill in Azure AI Search.
AnswerA

Custom NER can be trained with small datasets and improved with active learning.

Why this answer

Option C is correct because Custom NER allows you to train a model with a small dataset, especially if you use active learning to improve. Option A is wrong because the prebuilt entity recognition does not cover custom entities. Option B is wrong because the built-in extractor skills in AI Search are for indexing, not for custom entity extraction.

Option D is wrong because the Conversational PII entity extraction is for PII only.

896
MCQeasy

You need to analyze a live video stream from a security camera to detect people entering a restricted area. Which Azure AI service should you use?

A.Azure Video Indexer
B.Azure AI Custom Vision
C.Azure Video Analyzer for Media (deprecated)
D.Azure AI Face API
AnswerA

Azure Video Indexer supports live video analysis and can detect people and events in real time.

Why this answer

The correct answer is Azure Video Indexer (Video Analyzer) which supports live video analysis and can detect people and events. Option A is wrong because Video Analyzer for Media (now part of Video Indexer) is for analyzing pre-recorded video. Option B is wrong because Custom Vision is for image classification, not live video.

Option D is wrong because Face API is for facial recognition, not generic people detection in live video streams.

897
Multi-Selecteasy

Which TWO capabilities are available in Azure AI Search to improve search relevance? (Choose two.)

Select 2 answers
A.Filters
B.Indexers
C.Scoring profiles
D.Semantic ranking
E.Synonym maps
AnswersC, D

Scoring profiles boost results based on criteria.

Why this answer

Options A and D are correct. Scoring profiles allow boosting by field values or freshness. Semantic ranking re-ranks results to improve relevance.

Option B is wrong because synonym maps improve recall, not relevance. Option C is wrong because filters restrict results but do not improve relevance ranking. Option E is wrong because indexers are for data ingestion, not relevance.

898
Multi-Selectmedium

Which THREE practices should be followed to secure an Azure AI solution that uses Azure OpenAI Service and Azure AI Search?

Select 3 answers
A.Store API keys in Azure Key Vault but use them directly in application code.
B.Use managed identities to authenticate between Azure OpenAI and Azure AI Search.
C.Place all AI services in a DMZ subnet with public IP addresses.
D.Require that all client applications use HTTPS with TLS 1.2 or higher.
E.Enable firewall and private endpoints for all AI service endpoints.
AnswersB, D, E

Managed identities provide secure authentication without secrets.

Why this answer

Option B is correct because managed identities allow Azure OpenAI and Azure AI Search to authenticate securely without storing or rotating credentials. By assigning a managed identity to the Azure OpenAI service, you can grant it access to Azure AI Search via Azure RBAC, eliminating the need for API keys or shared access keys in code or configuration.

Exam trap

The trap here is that candidates often think storing keys in Key Vault is sufficient for security, but the question tests whether you understand that managed identities eliminate the need to handle keys altogether, and that public endpoints (even in a DMZ) are not secure for AI services.

899
MCQhard

Your company is using Azure OpenAI Service to generate marketing copy. The compliance team requires that all generated content be reviewed for sensitive data before delivery. You need to implement a solution that automatically scans the output for personally identifiable information (PII) and blocks it if detected. Which service should you integrate?

A.Azure AI Content Safety with a custom blocklist
B.Microsoft Purview Information Protection
C.Microsoft Defender for Cloud Apps
D.Azure AI Language PII detection
AnswerA

Content Safety can detect and block sensitive content including PII in real-time.

Why this answer

Option A is correct because Azure AI Content Safety with a custom blocklist can flag and block PII in generated content. Option B is wrong because Microsoft Purview Information Protection is for data classification and labeling, not real-time scanning of generative output. Option C is wrong because Azure AI Language PII detection is for recognizing PII but not built for blocking in a pipeline.

Option D is wrong because Microsoft Defender for Cloud Apps is for cloud app security, not content filtering.

900
MCQhard

You are using Azure AI Custom Vision to classify images of animals. The training set has 1000 images of cats and 1000 images of dogs. After training, the model performs well on the test set. However, when deployed, it misclassifies images of wolves as dogs. What is the most likely cause?

A.The training set does not include enough negative examples that look like dogs but are not.
B.The probability threshold is set too low.
C.The model is overfitted to the training data.
D.The training set has class imbalance.
AnswerA

Lack of hard negatives causes false positives.

Why this answer

Option B is correct because the training set likely lacks sufficient representative images of wolves, so the model learned that all large, wolf-like animals are dogs. Class imbalance (A) is not an issue here since both classes have 1000 images. Overfitting (C) would show poor performance on test set.

Probability threshold (D) is not the root cause.

Page 11

Page 12 of 14

Page 13