AWS Gen AI Practitioner Course Notes

Week 1

Pre-assessment Questions

1. What is difference between Machine Learning and Generative AI ?

2. Which AWS service allows you to choose from a broad range of foundation models from industry-leading providers ?

• Amazon Lex

• Amazon Bedrock

• Amazon Q

• Amazon SageMaker

• Amazon Lex : Building conversational interfaces and chatbots

• Amazon Q : AWS's AI-powered assistant for business productivity

• Amazon SageMaker : Builidng, training and deploying custom ML models

  • Prodvides some foundational model capabilities through SageMaker JumpStart

• Amazon Bedrock : Focusses on easy access to pre-trained foudation mdoels from various providers

1. Which is a key aspect of responsible AI that is especially important for generative AI?

• Reducing model size

• Avoiding toxicity

• Increasing model speed

• Improving model accuracy

1. Which techniques can be used to promote expected and trustworthy outputs from generative AI models? (Select THREE.)

• Employing human and AI-assisted reviews

• Automating the entire process

• Reducing training data

• Tailoring or filtering model output

• Increasing model complexity

• Developing responsible AI policies

1. Which is a key step in getting started with generative AI?

• Outsource all generative AI projects to external vendors

• Implement a full-scale generative AI solution

• Ignore organizational processes and focus only on technology

• Identify a generative AI use case

Module 1 : Introducing Generative AI

Terms and Definitions

• Artificial intelligence (AI):

  • refers to machines capable of performing tasks that typically require human intelligence.

• Machine learning (ML) :

  • is a subset of AI in which systems learn from structured data and improve performance of a specific task over time without being explicitly programmed to do so. It takes a set of inputs and predicts an output based on an algorithm trained with labelled data.

• Deep learning (DL) :

  • is a subset of machine learning that uses artificial neural networks to learn from unstructured data in a way that is inspired by the human brain. DL is capable of taking in very large amounts of unlabeled data, but its goal is still to make a specific prediction.

• Generative AI :

  • is a sub-field of AI. It’s a type of AI that can create new content and ideas, including conversations, stories, images, videos, and music

• Foundation Models

  • FMs are used by Generative AI

  • These are deep learning models that are trained on massive amounts of data and undergo extensive traning

  • Models might be trained for text, video, audio, image generation, customized outputs or multi-modal outputs such as financial models

  • Gen AI uses pre-trained FMs, making it attractive. It might have been challenging and resource-intensive using traditional ML approaces

• Large Language Models (LLMs)

• Prompts

• Parameters

• Prompt Engineering

• Augmenting

• Fine-tuning