What is generative AI and why it exploded in AI for Everyone - Practice Questions & Exercises

Challenge - 5 Problems

🧠 Conceptual

intermediate

What is the main function of generative AI?

Generative AI is designed to:

A Generate new content like text, images, or music

B Classify images into categories

C Detect anomalies in data

D Compress large datasets efficiently

💡 Hint

Think about what 'generative' means in everyday language.

Generative AI creates new content such as text, images, or music, unlike models that only classify or analyze existing data.

🧠 Conceptual

intermediate

Why did generative AI become very popular recently?

Which of these is NOT a main reason for the recent explosion of generative AI?

A Advances in computing power like GPUs

B Availability of large datasets for training

C Improved algorithms and model architectures

D Decrease in internet usage worldwide

💡 Hint

Consider what helps AI learn better and faster.

The explosion is due to better hardware, more data, and smarter algorithms. Internet usage decreasing would not help AI progress.

❓ Metrics

advanced

Which metric best measures the quality of text generated by a generative AI model?

You want to evaluate how well a generative AI writes sentences. Which metric is most suitable?

A Accuracy

B BLEU score

C Confusion Matrix

D Mean Squared Error

💡 Hint

Think about comparing generated text to reference text.

BLEU score measures how close generated text is to human-written text, unlike accuracy or error metrics used in other tasks.

❓ Model Choice

advanced

Which model architecture is commonly used for generative AI in text?

Which of these model types is most commonly used to generate human-like text?

A Transformers

B Recurrent Neural Networks (RNNs)

C Convolutional Neural Networks (CNNs)

D Support Vector Machines (SVMs)

💡 Hint

This architecture uses attention mechanisms and is behind popular models like GPT.

Transformers use attention to handle long-range dependencies and are state-of-the-art for text generation.

🔍 Analysis

expert

What error will this code cause when generating text with a transformer model?

Consider this Python snippet using a transformer model: ```python output = model.generate(input_ids, max_length=50, do_sample=True) print(output) ``` If `input_ids` is missing or None, what error will occur?

A ValueError: max_length must be positive

B AttributeError: 'NoneType' object has no attribute 'generate'

C ValueError: input_ids have to be passed and be a tensor or list

D RuntimeError: CUDA out of memory

💡 Hint

Think about what happens if you pass None where a list or tensor is expected.

If input_ids is None, the model.generate() method explicitly checks and raises a ValueError because input_ids must be provided as a tensor or list.