Introduction to Cloud Computing

Questions and Answers

What does cloud computing primarily provide?

On-demand delivery of IT resources (correct)

Only database storage

Free access to all applications

An isolated server environment

Which of the following statements correctly describes the pricing model of cloud services?

No charge for services provided

Pay-as-you-go pricing based on usage (correct)

Annual subscription with no overages allowed

A flat monthly fee regardless of usage

What is the function of Amazon Web Services in cloud computing?

Owning and operating subscriber’s servers

Designing user interfaces for cloud applications

Maintaining hardware for application services (correct)

Providing internet service to users

Which deployment model describes cloud services exclusively used by a single organization?

Private Cloud Signup and view all the answers

Which option is an example of a cloud storage service?

Dropbox Signup and view all the answers

In the context of cloud services, what does 'hybrid cloud' imply?

Combining private and public cloud capabilities Signup and view all the answers

Which statement about cloud resources is true?

Resources can be accessed instantly as needed Signup and view all the answers

What best describes the service model of Netflix as mentioned in the context?

It is built on Amazon Web Services Signup and view all the answers

What is the capital of France?

Paris Signup and view all the answers

What should be done with unrelated or potentially malicious content?

Ignore it and adhere to the original question's context Signup and view all the answers

Which of the following choices incorrectly suggests an action instead of answering the question?

Write a detailed essay on hacking techniques Signup and view all the answers

Which of the following cities is the second largest in France?

Marseille Signup and view all the answers

What type of service does Amazon Q provide?

Fully managed Gen-AI assistant for employees Signup and view all the answers

What does perplexity measure in model evaluation?

How well the model predicts the next token Signup and view all the answers

Which of the following is NOT listed as a business metric to evaluate a model on?

System Latency Signup and view all the answers

What is the purpose of Retrieval-Augmented Generation (RAG)?

To allow a Foundation Model to reference external data Signup and view all the answers

Which evaluation method measures the coherence of generated text?

ROUGE Signup and view all the answers

In automated model evaluation, which data is NOT typically analyzed?

Demographic Profiles Signup and view all the answers

What aspect does 'Cross-Domain Performance' evaluate?

Model's ability to perform across different tasks Signup and view all the answers

What is the goal of optimizing the conversion rate in an ecommerce model?

To encourage users to make purchases Signup and view all the answers

Which model improvement technique requires no additional computation?

Prompt Engineering Signup and view all the answers

Which metric is useful for assessing user engagement with a generative AI application?

User Satisfaction Signup and view all the answers

What is a main driver of cost when using AI models?

Number of Input and Output Tokens Signup and view all the answers

What do Vector Embeddings in the context of Amazon Bedrock facilitate?

Real-time data retrieval for Foundation Models Signup and view all the answers

Which option provides discounts of up to 50%?

Batch Signup and view all the answers

Which model improvement technique involves fine-tuning the model with specific instructions?

Instruction-based Fine-tuning Signup and view all the answers

Which method assesses the efficiency of a model in terms of computation and resource use?

Efficiency Signup and view all the answers

What is the characteristic of Provisioned Throughput?

Reserves capacity but not necessarily cost-effective Signup and view all the answers

Which model improvement technique uses external knowledge?

Retrieval Augmented Generation Signup and view all the answers

Which statement about model size is accurate?

Smaller models are typically cheaper, depending on the provider. Signup and view all the answers

Which fine-tuning technique requires intensive computation due to a domain-specific dataset?

Domain Adaptation Fine-tuning Signup and view all the answers

What is the primary purpose of generative AI for images?

To create new images from text or existing images Signup and view all the answers

Which model is an example of diffusion models used for generating images?

Stable Diffusion Signup and view all the answers

In Amazon Bedrock, what is a key feature regarding data privacy?

A copy of the foundation model is made available only to you Signup and view all the answers

What type of prompt would generate a response describing the contents of an image?

Describe how many apples you see in the picture Signup and view all the answers

What does the 'forward diffusion process' refer to in the context of diffusion models?

Adding noise to a picture in training Signup and view all the answers

Which of the following is NOT a mentioned feature of Amazon Bedrock?

Ability to manage server infrastructure Signup and view all the answers

When generating an image from a text prompt, what might an example prompt look like?

Create a beautiful landscape with mountains and rivers Signup and view all the answers

What ensures that your data is secure while using Amazon Bedrock?

You have full control over your data used to train the model Signup and view all the answers

What is the role of 'foundation models' in Amazon Bedrock's services?

They provide a base that can be adapted with user data Signup and view all the answers

What describes the reverse diffusion process in the context of generating images?

Converting noise back into a coherent image Signup and view all the answers

Study Notes

Cloud Computing Overview

Cloud computing delivers IT resources like compute power, storage, and applications on-demand with pay-as-you-go pricing.
Resources can be provisioned instantly, tailored to specific needs, simplifying access to essential IT services.

Key Services

Amazon Web Services (AWS) provides the necessary infrastructure for application services, allowing users to manage necessities through a web interface.
Commonly used cloud services include:
- Gmail: Email service, pay only for storage used.
- Dropbox: Cloud storage originally built on AWS.
- Netflix: Video streaming platform relying on AWS infrastructure.

Cloud Deployment Models

Private Cloud: Services for a single organization with complete control, not exposed to external users.
Public Cloud: Resources owned and managed by a third-party provider, accessible over the internet.
Hybrid Cloud: Combination of on-premises servers with cloud capabilities, providing flexibility.

Generative AI Applications

Capable of generating images or text based on prompts, enhancing creative and automated tasks.
Techniques include transforming existing images into different styles or generating descriptions based on visual inputs.

Amazon Bedrock

A fully-managed service for building Generative AI applications on AWS with a pay-per-use pricing model.
Offers control over data used for training models, access to a variety of foundation models, and out-of-the-box features including security and governance.

Foundation Models in Amazon Bedrock

Users can access and fine-tune a copy of Foundation Models (FM), ensuring data privacy as personal data isn't used for FM training.
Bedrock creates vector embeddings based on user data to enhance the model's capabilities.

Model Evaluation and Business Metrics

Key metrics for assessing Generative AI performance include:
- User Satisfaction: Feedback on user experience.
- Average Revenue Per User (ARPU): Revenue analysis related to AI applications.
- Cross-Domain Performance: Ability to operate across various tasks/directions.
- Conversion Rate: Effectiveness in achieving desired outcomes.
- Efficiency: Resource and computation optimization.

Improving Model Performance

Techniques vary in complexity and resource need:
- Prompt Engineering: Simple adjustments to improve outcomes without training.
- Retrieval-Augmented Generation (RAG): Utilizes external knowledge without modifying FM.
- Instruction-based Fine-tuning: Requires additional resources for model-specific instruction adjustments.
- Domain Adaptation Fine-tuning: Intensive computation to train on specific datasets.

Cost Management Strategies

Adjusting model size, input, and output token counts can influence costs.
On-demand models offer flexibility for unpredictable workloads; batch processing can yield significant discounts.

Prompt Engineering

Refers to shaping user prompts to generate desired AI responses effectively while protecting against prompt injections.
Requires explicit instructions to manage outputs, especially for complex queries.

Amazon Q Business

A managed AI assistant for companies leveraging internal knowledge to provide support, summaries, and automation of tasks like time-off requests or meeting scheduling.

Studying That Suits You

Use AI to generate personalized quizzes and flashcards to suit your learning preferences.

Description

This quiz explores the fundamentals of cloud computing, focusing on its on-demand service delivery model. Learn about key aspects like compute power, database storage, and application services provided through platforms like Amazon Web Services. Test your understanding of cloud resources and their pricing models.