Certificação Microsoft AI-900

Compartilhado por um usuário do Flashcards.gg · gerado com IA · Denunciar este conjunto

Estes cartões cobrem os conceitos do exame de certificação Microsoft AI-900 e os serviços de IA do Azure com cenários práticos de negócios. O conjunto destina-se a candidatos que se preparam para a prova.

50 cartões Português Nível: Iniciante IA e ciência de dados Publicado
Toque no cartão para virar

Cartões · 50

Azure AI Vision: Optical Character Recognition (OCR)
An accounting firm automatically digitizes paper receipts and extracts totals and merchant names into their ERP.
Azure AI Vision: Image Analysis (Tagging and Categorization)
An e-commerce retailer automatically tags uploaded product photos with descriptive keywords like dress, red, and summer.
Azure AI Vision: Object Detection
A retail supermarket monitors shelf cameras to identify when cereal boxes run out of stock and triggers restocking alerts.
Azure AI Vision: Image Classification
A manufacturing plant inspects factory parts on a conveyor belt to classify them as either defective or non-defective.
Azure AI Vision: Face API (Detection)
A photo management app detects human faces in family albums to let users frame or crop thumbnails automatically.
Azure AI Vision: Spatial Analysis
A transit authority analyzes security video feeds to monitor pedestrian foot traffic and ensure social distancing in stations.
Azure AI Document Intelligence: Prebuilt Receipt Model
An expense management app scans employee lunch receipts to extract merchant, date, tax, and total expense values instantly.
Azure AI Document Intelligence: Prebuilt Invoice Model
A corporate finance department parses PDF vendor invoices to extract line items, amounts, and bank IBANs without manual entry.
Azure AI Document Intelligence: Prebuilt ID Model
A car rental company scans customer driver licenses during mobile check-in to verify name, expiration date, and license number.
Azure AI Document Intelligence: Custom Extraction Model
A legal firm trains a model on standardized non-disclosure agreements to extract client-specific termination clauses automatically.
Azure AI Language: Sentiment Analysis
A telecommunications company monitors Twitter mentions to flag dissatisfied customers and route them to retention specialists.
Azure AI Language: Key Phrase Extraction
A hotel chain processes thousands of guest feedback forms to discover frequent recurring topics like breakfast and air conditioning.
Azure AI Language: Named Entity Recognition (NER)
A news aggregator identifies public figures, locations, and organizations mentioned in articles to create hyperlinked tags.
Azure AI Language: Personally Identifiable Information (PII) Detection
A hospital redacts patient Social Security numbers and phone numbers from doctor notes before sharing records with researchers.
Azure AI Language: Language Detection
A global support portal identifies incoming inquiry languages to direct tickets to appropriately qualified multilingual agents.
Azure AI Language: Conversational Language Understanding (CLU)
A smart home assistant identifies intent like TurnOnLight and extracts entities like LivingRoom from customer spoken requests.
Azure AI Language: Question Answering
A university creates a chatbot directly from its student handbook PDF to answer queries about registration deadlines 24/7.
Azure AI Language: Text Summarization
A financial research platform produces concise bullet-point summaries of hundred-page corporate annual earnings reports.
Azure AI Speech: Speech-to-Text
A call center transcribes daily customer service phone recordings into text files for compliance auditing and analytics.
Azure AI Speech: Text-to-Speech
An audiobook publisher converts manuscript text files into natural-sounding audio narrations using neural voices.
Azure AI Speech: Speech Translation
An international webinar tool translates a keynote speaker spoken English into real-time audio streams in Japanese and Spanish.
Azure AI Speech: Speaker Recognition
A telephone banking system authenticates customers by matching their voice passphrase against an enrolled voice profile.
Azure AI Translator: Text Translation
An international retail website translates English product descriptions into Portuguese and French for regional shoppers.
Azure AI Translator: Document Translation
A multinational firm translates user manuals from English to German while preserving formatting, tables, and images.
Azure AI Search: Cognitive Search Indexing
A law firm searches millions of scanned legal briefs and contracts using OCR and key phrase enrichment to find case precedents.
Azure AI Search: Semantic Ranking
An internal knowledge base ranks HR policy search results by user intent rather than simple keyword matches.
Azure AI Search: Vector Search
An online clothing store allows users to search items matching concepts like lightweight jacket for rainy hiking.
Azure OpenAI Service: GPT-4 Text Generation
A marketing agency drafts initial personalized email campaigns and social media posts based on product release notes.
Azure OpenAI Service: Embeddings
A real estate portal converts property listings into numerical vectors to recommend houses with similar descriptions.
Azure OpenAI Service: DALL-E
A video game studio generates concept art variations and environmental textures rapidly from text prompts.
Azure Machine Learning: Automated ML (AutoML)
A bank finds the best classification algorithm and hyperparameters to predict credit card churn without manual model coding.
Azure Machine Learning: Designer
A business analyst builds a customer lifetime value prediction pipeline by dragging and dropping data pre-processing modules.
Azure Machine Learning: Managed Endpoints
A fraud detection team deploys a trained ML model behind a real-time REST API to score card transactions in milliseconds.
Machine Learning: Supervised Regression
A real estate agency predicts continuous home selling prices based on square footage, location, and number of bedrooms.
Machine Learning: Supervised Binary Classification
An email provider builds an algorithm to label incoming messages as either Spam or Not Spam.
Machine Learning: Supervised Multiclass Classification
A logistics company classifies customer feedback into three categories: Delivery, Billing, or Customer Service.
Machine Learning: Unsupervised Clustering
A supermarket segments loyalty members into distinct shopper clusters without prior labels to tailor discount coupons.
Machine Learning: Anomaly Detection
An industrial IoT platform detects unusual pressure spikes in gas pipelines to alert operators before failure occurs.
Responsible AI: Fairness
A loan approval system is audited to ensure credit approval rates do not discriminate against applicants by gender or race.
Responsible AI: Reliability and Safety
An autonomous vehicle AI undergoes rigorous simulation testing to guarantee predictable braking in heavy fog and rain.
Responsible AI: Privacy and Security
A healthcare AI platform ensures patient diagnostic data is encrypted at rest and in transit during model training.
Responsible AI: Inclusiveness
A screen reader service provides speech synthesis and image descriptions so visually impaired users can navigate web forms.
Responsible AI: Transparency
An insurance assessment tool provides explanations of why a particular policy claim payout was reduced or denied.
Responsible AI: Accountability
A hospital designates human clinical review boards to remain legally and ethically responsible for diagnostic AI tools.
Azure Bot Service
An airline integrates a customer service bot across Web Chat, Microsoft Teams, and WhatsApp to handle flight rebookings.
Azure AI Content Safety: Text Moderation
A gaming platform filters out hate speech and profanity in multiplayer public text chats in real time.
Azure AI Content Safety: Image Moderation
A social network automatically scans and blocks user-uploaded images containing graphic violence or explicit content.
Retrieval-Augmented Generation (RAG)
An enterprise chatbot retrieves internal SharePoint manuals to supply OpenAI with context for accurate company-specific answers.
Azure AI Studio
An enterprise engineering team develops, evaluates, and deploys customized generative AI copilot applications securely.
Evaluation Metric: Precision and Recall
A cancer diagnostic lab optimizes recall to avoid false negatives while balancing precision to prevent unnecessary biopsies.