Quiz 2: Advanced Topics – Reinforcement Learning, Transformers,

What is a reward in reinforcement learning?

• An action taken by the agent

• The final output of the model

• Feedback given to the agent

• Data preprocessing step

Answer explanation

A reward guides the agent’s behavior by signaling success or failure.

Which of these is an application of reinforcement learning?

• Controlling a robotic arm

• Predicting stock prices

• Image recognition

• Clustering customer data

Answer explanation

Reinforcement learning is ideal for sequential decision-making tasks like robotics.

What is the fundamental concept behind transformer architectures?

• Attention mechanism

• Data clustering

• Tree-based structure

• Sequence alignment

Answer explanation

Attention mechanisms enable transformers to weigh the importance of each input token.

What is the purpose of diffusion models in AI?

• Predicting numerical values

• Generating high-quality images

• Simplifying datasets

• Data preprocessing

Answer explanation

Diffusion models generate realistic data like images or audio by modeling complex distributions.

Which generative model uses both images and text as inputs?

• GPT

• Stable Diffusion

• Vision-Language Models

• Mid-Journey

Answer explanation

Vision-language models process both image and text inputs to generate relevant outputs, such as captions.

What is the main advantage of GPT-based models?

• Simplicity in architecture

• High accuracy in numerical data

• Language generation capabilities

Answer explanation

GPT models excel in generating coherent and contextually relevant text.

Which model is primarily used for language translation?

• Regression model

• Transformer model

• K-Means clustering

• Random Forest

Answer explanation

Transformer models power translation systems like Google Translate by processing sequences efficiently.

What is the role of reinforcement learning in robotics?

• Learning to maximize task rewards

• Clustering similar images

• Predicting continuous data

• Building models from text

Answer explanation

In robotics, reinforcement learning helps optimize sequences of actions for tasks.

Which of these is a key component of transformer architecture?

• Convolution layers

• Recurrent layers

• Attention mechanism

• Gradient descent

Answer explanation

The attention mechanism allows transformers to focus on relevant parts of the input sequence.

What is a common application of diffusion models?

• Spam filtering

• Autonomous vehicles

• Image generation

• Predicting numerical outputs

Answer explanation

Diffusion models are widely used for generating high-quality, realistic images.