Generative AI Intro

Foundational Concepts

• Artificial Intelligence (AI): The broad field where computers mimic "intelligent" human behaviors

• Machine Learning (ML): A subset of AI where systems learn patterns from data without explicit programming.

• Generative AI: A type of AI that creates new content (text, images, code, etc.) rather than just analyzing existing data.

Key Model Types

• Large Language Models (LLMs): Powerful AI models trained on massive amounts of text data. They are the backbone of many generative AI applications

• Generative Pre-trained Transformer (GPT): A popular architecture for LLMs, known for producing realistic and coherent text.

• Diffusion Models: Used especially for image generation. They "denoise" random patterns to gradually create images.

Important Techniques

• Prompting: How you ask the AI to generate something. This includes the structure and style of your input.

• Fine-tuning: Adapting a pre-trained model to perform a specific task by using additional, smaller amounts of targeted data.

• Embedding: Converting words or images to numerical representations the computer understands, allowing analysis and manipulation.

Challenges & Considerations

• Hallucination: When AI generates false or misleading information.

• Bias: When AI reflects prejudices in its training data which can lead to problematic outputs.

• Explainability: The level to which we can understand how the AI reaches its outputs. This is important for trust and fairness.

"Getting started with artificial intelligence" prompt summary. Gemini, 4 April, Google, gemini.google.com.