Microsoft Azure AI Fundamentals Flashcards

An NLP technique that identifies the main talking points or salient phrases in a body of text, summarizing what it is about.

The process of converting spoken audio into written text transcription.

The process of converting written text into spoken audio output.

700, reported on a scaled score range of 1 to 1,000.

Software that imitates human behaviors and capabilities such as perceiving, reasoning, learning, and making decisions or predictions from data.

A subset of AI in which models learn patterns from historical data and use those patterns to make predictions on new, unseen data, rather than following explicit hand-coded rules.

Regression predicts a continuous numeric value (e.g., a price), while classification predicts a discrete category or class label (e.g., spam vs. not spam).

Supervised learning trains on labeled data (inputs paired with known outputs), while unsupervised learning finds structure or patterns in unlabeled data.

The known output value associated with a set of features in training data, which the model learns to predict.

An AI discipline that enables software to interpret and understand visual input from images or video, such as identifying objects, faces, or text.

Image classification assigns one or more labels to an entire image, while object detection identifies and locates multiple individual objects within an image using bounding boxes.

A computer vision capability that extracts printed or handwritten text from images or scanned documents into machine-readable text.

Facial detection locates the presence and position of faces in an image, while facial recognition identifies or verifies a specific individual's identity from a face.

An AI field focused on enabling computers to understand, interpret, and generate human language, both written and spoken.

An NLP technique that determines the emotional tone (positive, negative, neutral) expressed in a piece of text.

An NLP task that identifies and categorizes key elements in text, such as people, places, organizations, and dates.

AI that creates new original content — such as text, images, code, or audio — by learning patterns from large datasets, rather than simply classifying or analyzing existing content.

A deep learning model trained on massive text datasets that can understand and generate human-like language for tasks like completion, summarization, and conversation.

The input text or instruction given to a generative AI model to guide the content it produces.

The practice of designing and refining prompts to elicit more accurate, relevant, or useful responses from a generative AI model.

A pattern that grounds a generative AI model's responses by retrieving relevant information from an external knowledge source (such as your own documents) and injecting it into the prompt before generation.

Supplying a model with authoritative, contextual data (e.g., via RAG) so its outputs are based on relevant facts rather than only on its pretrained knowledge, reducing hallucination.

A confident but factually incorrect or fabricated response generated by an AI model that is not grounded in real data.

An AI system that can autonomously plan, make decisions, and take multi-step actions — often invoking tools or other services — to accomplish a goal with limited human intervention.

Microsoft's unified platform for building, customizing, evaluating, and deploying AI solutions and agents, providing access to a model catalog and orchestration tooling.

Fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability.

Ensuring AI systems treat all people equitably and do not create or reinforce bias against particular groups.

A curated collection of information (documents, FAQs) that a language model or bot draws on to ground its answers to user queries.

AI capable of processing and reasoning across multiple input types simultaneously, such as text, images, and audio, rather than being limited to a single modality.