Flashcards: Artificial Intelligence Fundamentals — 43 cards

What is artificial intelligence in computer science?

A branch aiming to build systems performing tasks requiring human intelligence.

What do most current AI systems specialize in?

Performing one task or a limited group of tasks.

Name one example of an AI application.

Spam filtering.

What is the broadest field among AI, machine learning, and deep learning?

Artificial intelligence is the broadest field.

Which field is a part of artificial intelligence?

Machine learning is a part of artificial intelligence.

What specialized approach within machine learning does generative AI commonly use?

Generative AI commonly relies on deep-learning models.

What is machine learning?

Machine learning is a method where systems learn patterns from data to make predictions or judgments.

What characterizes deep learning within machine learning?

Deep learning uses neural networks to learn complex patterns and relies on large amounts of data.

What does generative AI do with deep learning?

Generative AI uses deep learning to generate entirely new content and data.

What does a machine learning system study to make predictions?

Many previous examples.

What does a machine learning system discover from previous examples?

Recurring patterns.

What does a machine learning system use to predict new cases?

Discovered recurring patterns.

What does a spam filter learn from?

Previous email messages.

What does a spam filter identify to classify emails?

Recurring patterns.

What does a spam filter predict for a new message?

Whether it is spam.

From what data does a recommendation system learn?

Previous purchasing data.

What does a recommendation system use to suggest new products?

Detected patterns from past data.

What is an artificial neural network inspired by?

Connections between neurons.

How does an artificial neural network process inputs?

Using connected layers to produce outputs.

What does a neural network analyze in visual recognition?

Visual patterns like lines, colors, and shapes.

What is the final output of a neural network in visual recognition?

A recognition or decision.

When do deep-learning systems tend to make mistakes?

With too few examples or new patterns differing from training data.

How does a traditional classification system differ from a generative system?

A traditional system selects predefined options, a generative system creates new content.

What initiates a generative AI interaction?

A user prompt initiates a generative AI interaction.

What is the output of a generative AI after applying a trained model?

Generated text, images, audio, or other content.

What can ChatGPT generate from a prompt about a robot?

A completely new short story about a robot.

Name three types of content generative AI can produce.

Texts, images, and music.

Can generative AI create programming code and translations?

Yes, it can produce programming code and translations.

Is generative AI capable of producing videos and animations?

Yes, it can produce videos and animations.

Which tasks require artificial intelligence in the given examples?

Speech recognition, image recognition, translation, and recommendations require artificial intelligence.

Which tasks are considered simple computational tasks in the examples?

A simple calculator and a remaining-time display are simple computational tasks.

How are applications classified according to their primary function?

Applications are classified by primary function into machine learning, deep learning, and generative AI.

What is the primary focus of machine learning applications?

Machine learning focuses on learning patterns for prediction.

What does deep learning primarily focus on?

Deep learning focuses on complex neural-network processing.

What is the main function of generative AI applications?

Generative AI focuses on creating new content.

What is a hallucination in AI systems?

An answer that seems convincing but is invented or unsupported.

Why might AI systems lack information on current events?

Their knowledge can stop at a particular date.

What can AI outputs reflect from their training data?

Bias or errors present in the data.

Is an AI system a final source of truth without human review?

No, it requires human review to be a final source.

How can you verify AI output effectively?

By checking original sources, comparing with reliable references, testing logic, and taking personal responsibility.

How should generative AI be used in relation to thinking and writing code?

As an aid for understanding and thinking, not as a substitute or complete replacement.

What is the main benefit of generative AI?

Speed in producing large amounts of content and supporting brainstorming.

What are the main risks of using generative AI?

Hallucinated information, difficulty distinguishing fact from imagination, and weakened critical thinking.