Practical TypeScript Artificial Intelligence Programming

Cover Material, Copyright, and License

Preface

1. About the Author
2. Using the Example Code
3. NPM Security Concerns
4. Acknowledgements

Setting Up a TypeScript Development Environment

1. Installing Node.js
2. Installing TypeScript and tsx
3. Creating a New Project
4. Running Existing Projects in directory source-code
5. Running TypeScript Files
6. Environment Variables
7. Code Formatting (Optional)

A TypeScript Tutorial for Command-Line AI Programs

1. Type Basics
2. Interfaces and Type Aliases
3. Functions
4. Async/Await and Promises
5. Classes
6. Modules and Imports
7. Error Handling
8. Working with Files
9. Enums and Literal Types
10. Map, Set, and Iterators
11. Practical Patterns for AI Code
12. TypeScript Tutorial Wrap-up

Part 1 - Machine Learning

“Classic” Machine Learning

1. Example Material
2. Loading CSV Data
3. Classification Using K-Nearest Neighbors
4. Classic Machine Learning Wrap-up

Regression and Clustering

1. Regression: Predicting Housing Prices
2. Clustering: Discovering Groups in Data
3. Regression and Clustering Wrap-up

Exploratory Data Analysis and Feature Engineering

1. Exploratory Data Analysis
2. Feature Engineering
3. EDA and Feature Engineering Wrap-up

Anomaly Detection

1. What Is a Gaussian Distribution?
2. How the Detector Works
3. The Wisconsin Breast Cancer Dataset
4. Project Structure
5. Walking Through the Code
6. Running the Example
7. Using the API in Your Own Code
8. Understanding the Evaluation Metrics
9. Wrap Up

Part 2 - Deep Learning

The Basics of Deep Learning

1. Using TensorFlow.js for Building a Cancer Prediction Model

Natural Language Processing Using Deep Learning

1. Hugging Face and the Transformers.js Library
2. Comparing Sentences for Similarity Using Transformer Models
3. Deep Learning Natural Language Processing Wrap-up

Part IV - Overviews of Image Generation, Reinforcement Learning, and Recommendation Systems

Overview of Image Generation

1. Image Generation Using the Hugging Face Inference API
2. Image Generation Using Local Ollama Models
3. Recommended Reading for Image Generation

Overview of Reinforcement Learning (Optional Material)

1. Overview
2. An Introduction to Markov Decision Process
3. A Concrete Example: Q-Learning
4. Reinforcement Learning Wrap-up

Overview of Recommendation Systems (Optional Material)

1. TensorFlow Recommenders
2. Project Structure
3. Item-Based Collaborative Filtering
4. Embedding Matrix Factorization
5. Running the Examples
6. Comparing the Two Approaches
7. Using the API in Your Own Code
8. Recommendation Systems Wrap-up

Part 3 - Large Language Models

Introduction to Transformers and Large Language Models

1. The Transformer Architecture
2. Tokenization
3. From Transformers to Large Language Models
4. Key Capabilities of Modern LLMs
5. Practical Considerations

LLMs with Public APIs

1. Setup and Authentication
2. Text Generation
3. Thinking Models
4. Multi-Turn Conversations
5. Multimodal Input: Analyzing Images
6. Structured Output
7. Tool Use (Function Calling)
8. Practical Considerations
9. Summary

LLMs with Local Models

1. Installing Ollama
2. Downloading and Running Models
3. Using Ollama from TypeScript
4. Reasoning with Local Models
5. Conversation Memory with Ollama
6. Describe Content of Images
7. Adding Web Search Tools
8. OpenAI-Compatible API
9. Alternative Tools for Running Local Models
10. Hardware Considerations
11. Summary

An AI Command-Line Tool with Search Grounding and Persistent Cache

1. How It Works
2. Prerequisites
3. Project Structure
4. Keyword Extraction
5. The Cache Engine
6. The Main REPL Application
7. Running the Tool
8. Example Session
9. REPL Command Reference
10. Key Takeaways

Part 4 - Symbolic AI and Knowledge Representation

Classic Graph Search

1. Graphs and Search Representation
2. Implementing Graph Search in TypeScript
3. Example Run
4. Wrap-up

Chess Game with Alpha-Beta Search

1. How a Chess Engine Works
2. Project Structure
3. Walking Through the Code
4. Running the Example
5. Wrap Up

Part 5 - Knowledge Representation

Getting Setup To Use Graph and Relational Databases

1. Querying Wikidata with SPARQL and TypeScript
2. The SQLite Relational Database for Knowledge Representation

Optional Material: A Deeper Dive Into Semantic Web and Linked Data

1. Overview and Theory

Open Knowledge Format (OKF) Bundle Explorer

1. Inspiration and the Specification
2. The Knowledge Bundle Structure
3. Defining the OKF Data Model
4. Implementing the Knowledge Bundle
5. Building the Consumption Agent
6. Tying It All Together
7. Running the Explorer and Sample Output
8. Summary and Future Improvements