Generative AI Quiz – MCQ Questions and Answers
What is the primary function of AI in precision agriculture?
To create giant crops
To optimize crop yields and resource use
To replace farmers
To control weather patterns
What is the main drawback of using "greedy decoding" for text generation in LLMs?
Generated text lacks diversity
High computational cost
Inability to handle out-of-vocabulary words
Difficulty in controlling the output length
Which of the following is a potential issue with LLMs?
Generating biased or offensive content
Lack of common sense understanding
Difficulty in controlling the generated output
All of the above
What is the primary innovation of the Switch Transformer in large-scale language modeling?
Conditional computation using sparsely-activated experts
Increased model depth
Use of quantization techniques
Implementation of federated learning
What is backpropagation?
Propagating AI backwards
An algorithm for training neural networks
Backing up AI propagation
Propagating data back to its source
What is the key innovation of the Stable Diffusion model?
Improved text processing
Enhanced speech recognition
Implementation of quantum algorithms
High-quality image generation from text
What is a conditional GAN?
A GAN that generates samples based on class labels
A GAN with multiple generators
A GAN with multiple discriminators
A GAN that uses convolutional layers
What is the main goal of AI ethics?
To make AI more ethical
To ensure responsible AI development
To generate ethical guidelines
To classify ethical dilemmas
Which architecture introduced the concept of "inception modules"?
ResNet
VGG
GoogLeNet
AlexNet
Were the form elements recognized with greater than 70% confidence according to the response?
Yes
No
Confidence not specified
Varies by element
What does "one-shot" prompting refer to in the context of large language models?
Using a single example to guide the model's output
Generating only one response
Training the model with one data point
Running the model only once
Which of the following is a key challenge in using generative AI for language model fine-tuning?
Catastrophic forgetting
Handling different languages
Generating diverse outputs
All of the above
In prompt engineering, what is the "instruction-following" technique?
Training models to follow specific instructions
Providing step-by-step instructions in the prompt
Using instructional videos for AI training
Optimizing prompt length
Which architecture introduced the concept of "inverted residuals"?
VGG
AlexNet
MobileNetV2
Inception
How do generative AI models differ from traditional chatbots?
Generative AI can create new content, chatbots follow scripts
Generative AI is always more accurate
Chatbots are faster
Generative AI doesn't use language processing
What is the main application of AI in the gardening industry?
To replace gardeners
For plant identification and care recommendations
To control plant growth
To manufacture gardening tools
What is the main purpose of the Inception Score (IS) in evaluating GANs?
To measure the quality and diversity of generated samples
To speed up training
To reduce model complexity
To compress input data
In computer vision, what is the main purpose of panoptic segmentation?
To classify entire images
To detect object boundaries
To combine instance and semantic segmentation
To generate synthetic images
What is the main purpose of using generative AI in financial fraud detection?
To replace human auditors
To generate synthetic fraud scenarios for testing
To automatically approve all transactions
To predict stock market trends
What is "nucleus sampling" in the context of language models?
A technique for data augmentation
A method for model compression
A strategy for diverse text generation
A way to evaluate model performance
Score: 0/20