Generative AI Fundamentals Quiz Quiz

Test your knowledge of generative AI concepts, terminology, and key technologies with these easy multiple-choice questions.

1. Definition of Generative AI

   Which statement best describes what generative AI does?

   1. It creates new data such as text, images, or audio by learning patterns from training data.
   2. It only sorts and classifies data into categories.
   3. It deletes unnecessary data from large datasets.
   4. It performs manual calculations for researchers.
   5. It encrypts user information for security.

2. Difference From Traditional AI

   How is generative AI different from traditional predictive AI approaches?

   1. Generative AI works offline, but traditional AI is only online.
   2. Generative AI is used to create new content, while traditional AI focuses on prediction and classification.
   3. Generative AI updates databases directly, while traditional AI does not.
   4. Traditional AI requires no training data, unlike generative AI.
   5. Traditional AI only generates images, while generative AI creates code.

3. Popular Applications

   Which of the following is a common application of generative AI?

   1. Managing spreadsheet formulas automatically.
   2. Sorting physical mail by address.
   3. Natural language text generation, such as creating chatbot responses.
   4. Locating lost phones using GPS.
   5. Compressing video files for storage.

4. Understanding GANs

   What are Generative Adversarial Networks (GANs) made up of?

   1. A single linear regression model.
   2. A generator and a discriminator network that compete to improve generated data.
   3. A decision tree and a classifier.
   4. Blocks of handwritten rules.
   5. Two identical neural networks that never change.

5. GANs vs VAEs

   How does a GAN differ from a Variational Autoencoder (VAE) when generating data?

   1. GANs cannot process images but VAEs can.
   2. A GAN uses a competing generator and discriminator, while a VAE uses probabilistic encoding and decoding.
   3. A VAE is faster because it skips the encoding step.
   4. VAEs do not require training on any data.
   5. A GAN uses only random guessing, while a VAE does not.

6. Parts of a GAN

   In the context of generative adversarial networks, what does the generator component do?

   1. It deletes unusable information from inputs.
   2. It only labels images without changes.
   3. It creates synthetic data that mimics real data.
   4. It organizes files alphabetically.
   5. It measures the speed of network connections.

7. Prompt Engineering

   What is prompt engineering in large language models?

   1. It's optimizing internet connectivity for the model.
   2. It refers to repairing computer hardware parts.
   3. It means installing updates automatically.
   4. It's coding with binary instructions only.
   5. It's the practice of designing effective inputs to get desired responses from AI models.

8. GPT vs BERT

   Which key feature distinguishes GPT from BERT among language models?

   1. BERT is used for sorting emails, whereas GPT is not.
   2. GPT can only answer numeric questions.
   3. BERT is incapable of being trained on any data.
   4. GPT uses only images, but BERT uses only audio.
   5. GPT is unidirectional and suited for text generation, while BERT is bidirectional for understanding tasks.

9. Transformers Concept

   What is the main mechanism that enables transformers to process language sequences effectively?

   1. Transformers rely on random guessing for every step.
   2. Transformers require programming in only one language.
   3. Self-attention allows transformers to consider relationships between all parts of a sequence.
   4. They delete irrelevant data without any comparison.
   5. They use only fixed memory addresses.

10. Role of Loss Function

    What is the role of the loss function in training generative AI models?

    1. It creates unique neural network layers.
    2. It measures how accurate the model's output is and helps optimize its predictions.
    3. It encrypts the model's parameters.
    4. It deletes old training data regularly.
    5. It changes user privacy settings automatically.