Overview
This lecture provides an overview of the complex AI landscape, explaining key domains, how they overlap, and the basis for categorizing them.
Ways to Segment AI Domains
- AI domains can be defined by input type, output type, or model type.
- Natural Language Processing (NLP) focuses on processing human language inputs like voice or text.
- Generative AI (GenAI) focuses on generating new content—text, images, or audio.
- Domains often overlap, such as ChatGPT which uses both NLP and generative AI.
Model Types in AI
- Rule-based AI uses if-then logic rules created by humans for decision-making.
- Rule-based AI is limited because humans must know and define all rules, which is hard in changing scenarios.
- Machine Learning (ML) allows machines to determine optimal rules by using data-driven training.
- ML models are built in three steps: humans choose inputs and form, the machine optimizes parameters through training.
- Various mathematical forms in ML include straight lines, curves, trees, and neural networks.
Neural Networks & Deep Learning
- Neural networks mimic the human brain using interconnected nodes (neurons) in layers.
- Deep neural networks have more than two hidden layers and are called deep learning models.
- Deep learning is a subset of machine learning focused on complex models like large language models (LLMs).
- Advanced AI models such as LLMs and image generators are built using deep learning.
Model Training
- All machine learning models, including deep learning models, require a training process to learn from data.
- The lecture mentions that training methods will be discussed in the next chapter.
Key Terms & Definitions
- Natural Language Processing (NLP) — AI domain focused on understanding and processing human language inputs.
- Generative AI — AI that creates new text, image, or audio content.
- Rule-based AI — AI using explicitly coded if-then rules for decision-making.
- Machine Learning (ML) — AI domain where models learn rules from data via training.
- Neural Network — AI model inspired by the human brain, consisting of layers of nodes (neurons).
- Deep Learning — Subset of ML using neural networks with multiple hidden layers for complex tasks.
- Large Language Model (LLM) — An advanced AI model built using deep learning to process and generate human language.
Action Items / Next Steps
- Review the next chapter for details on model training methods.