Generative AI Essentials: Easy Interview Practice Quiz Quiz

1. Fundamentals of Generative AI

   Which of the following best describes what Generative AI does?

   1. A. It creates new data that resembles examples it has seen during training.
   2. B. It only classifies images into categories.
   3. C. It deletes unimportant features from data.
   4. D. It measures the speed of a computer program.
   5. E. It prevents networks from overfitting.

2. Common Model Types

   Which type of neural network is commonly used in image synthesis by Generative AI systems?

   1. A. Recurrent Neural Network (RNN)
   2. B. Generative Adversarial Network (GAN)
   3. C. Decision Tree
   4. D. Naive Bayes Model
   5. E. Random Forest

3. Text Generation Example

   If an AI model produces a new poem after reading many poems, which task is it performing?

   1. A. Summarization
   2. B. Clustering
   3. C. Generation
   4. D. Translation
   5. E. Segmentation

4. Key Transformer Component

   What main mechanism in Transformer models allows them to focus on relevant words when processing a sentence?

   1. A. Pooling
   2. B. Activation function
   3. C. Attention
   4. D. Dropout
   5. E. Regression

5. Training Technique

   Transfer learning, often used in Generative AI, involves which of the following?

   1. A. Training a brand new model from scratch for each task
   2. B. Deleting previous model weights before training
   3. C. Using a pre-trained model as a starting point for a new problem
   4. D. Generating random data and using it to train the model
   5. E. Only using labels with typos for learning

6. Real-World Applications

   Which is a common application of Generative AI in the music industry?

   1. A. Detecting plagiarism in songs
   2. B. Classifying genres based on lyrics
   3. C. Composing original melodies based on learned patterns
   4. D. Correcting spelling errors in band names
   5. E. Storing song lyrics in a database

7. Ethical Considerations

   One ethical concern with Generative AI is 'hallucination.' What does this refer to?

   1. A. The model plays music too loudly
   2. B. The model generates information that is factually incorrect
   3. C. The model falls asleep during training
   4. D. The model creates images without colors
   5. E. The model only works at night

8. Popular Architectures

   Which architecture is most commonly used for large language models, such as ones that can generate long text passages?

   1. A. Support Vector Machine (SVM)
   2. B. Transformer
   3. C. K-Means
   4. D. Linear Regression
   5. E. Feedforwad Network

9. Optimization Techniques

   Hyperparameter tuning in model training refers to which activity?

   1. A. Adjusting the input data to be larger than memory allows
   2. B. Randomly shuffling the layers in a network
   3. C. Selecting the best values for model settings like learning rate and batch size
   4. D. Turning off all model parameters before evaluation
   5. E. Choosing labels with intentional misspellings

10. Bias in AI Models

    If a language model generates stereotypical responses about a particular group, which issue does this demonstrate?

    1. A. Overfitting
    2. B. Bias
    3. C. Underfitting
    4. D. Regularization
    5. E. Hyperparmetrization