AI Image Generation Quiz

What are the three key technologies that underlie modern image-generation AI tools like Midjourney, DALL-E, and Stable Diffusion?

• Generative Adversarial Networks, Diffusion Models, and Transformer Models

• CNN, RNN, and LSTM networks

• Vision Models, Language Models, and Hybrid Models

• Neural Networks, Machine Learning, and Deep Learning

Which of the following best describes how Diffusion Models work?

• They use attention mechanisms to translate text into images

• They learn through competition between generator and discriminator networks

• They start with noise and gradually remove it to create an image

• They blend multiple existing images to create a new composition

How do Transformer Models in image generation differ from GANs?

• They process more images but with less detail

• They use attention mechanisms to understand relationships between text and images

• They only create black and white images

• They require manual adjustment of all parameters

What important element should be included when crafting prompts for image generation AI?

• Only the subject description

• Only the desired style

• Multiple specific details including lighting, style, perspective, and subject

• As few words as possible

Which AI image generation capability allows users to extend an existing image beyond its original boundaries?

• Inpainting

• Outpainting

• Style transfer

• Resolution upscaling

According to the course materials, what is one of the main advantages of Multimodal LLMs like DALL-E 3 and Gemini over dedicated image models?

• They produce higher resolution images

• They can maintain context during image creation and refinement

• They always produce better quality images

• They work faster than dedicated image models

In the historical timeline of generative art, which technology was highlighted as being used in the 1950s by Ben Laposky?

• Digital computers

• Oscilloscopes

• GANs

• Diffusion models