Introduction to Generative AI

Course Overview

• Instructor: Roger Martinez, Developer Relations Engineer at Google Cloud
• Objectives:
  • Define generative AI
  • Explain how generative AI works
  • Describe generative AI model types
  • Describe generative AI applications

What is Generative AI?

• Generative AI is a type of artificial intelligence technology that produces content such as:
  • Text
  • Imagery
  • Audio
  • Synthetic data

Context: Artificial Intelligence (AI)

• AI as a discipline in computer science focused on creating intelligent agents.
• AI encompasses methods to build machines that can reason, learn, and act autonomously.

Machine Learning (ML)

• Subfield of AI: Trains models from input data to make predictions.
• Key Concepts:
  • Supervised Learning:
    • Uses labeled data to make predictions.
    • Example: Predicting tip amounts based on historical data.
  • Unsupervised Learning:
    • Uses unlabeled data to discover patterns or groupings.
    • Example: Clustering employees based on tenure and income.

Deep Learning

• Subset of ML: Uses artificial neural networks to process complex patterns.
• Neural Networks:
  • Inspired by the human brain.
  • Can use both labeled and unlabeled data (semi-supervised learning).

Generative AI

• A subset of deep learning that can process labeled and unlabeled data using various methods.
• Model Types:
  • Generative Models: Generate new data based on learned distributions.
  • Discriminative Models: Classify or predict labels based on input data.

Distinguishing Generative AI

• Is it Generative AI?
  • Not GenAI: Output is a number, class, or probability.
  • Is GenAI: Output is natural language, audio, or images.

Mathematical Perspective

• Model Output: Y = f(X)
  • Y: dependent output
  • f: function used for prediction
  • X: input data
• Distinction between generative and traditional ML outputs.

Generative AI Process

• Involves training with code, labeled data, and unlabeled data to create a foundation model.
• Capable of generating various content types (text, images, audio, video).

Applications of Generative AI

• Can answer queries and generate content based on prompts.
• Example: Gemini AI model for text and image generation.

Types of Generative AI Models

1. Text-to-Text: Translates or manipulates text.
2. Text-to-Image: Generates images from text descriptions.
3. Text-to-Video: Produces video content from text input.
4. Text-to-3D: Creates 3D models based on text descriptions.
5. Text-to-Task: Performs specific tasks based on text input.

Foundation Models

• Large, pre-trained models adaptable to various tasks (e.g., sentiment analysis, image recognition).
• Vertex AI Model Garden: Resource for accessing foundation models in Google Cloud.

Generative AI in Coding

• Code Generation Example: Converting Python to JSON using Gemini.
• Skills: Debugging, explaining code, crafting SQL queries.

Google Cloud Tools for Generative AI

1. Vertex AI Studio: Customizes generative AI models easily.
2. Vertex AI: Builds chatbots and digital assistants with minimal coding.
3. Gemini Model: Multimodal capabilities beyond text understanding.

Conclusion

• Summary of generative AI principles and applications.
• Encouragement to explore further learning resources.