Where to Find Beginner-Friendly Example Scripts in AI for Beginners: A Complete Guide

All beginner-friendly example scripts in the Microsoft AI for Beginners repository are located in the top-level examples/ directory, with additional runnable notebooks available throughout the lessons/ folder.

The microsoft/AI-For-Beginners repository provides a curated collection of self-contained Python scripts designed specifically for newcomers to artificial intelligence. These beginner-friendly example scripts demonstrate core AI concepts without requiring complex framework setup or deep theoretical background. Each file in the examples/ directory can be executed directly with python <script>.py to see immediate results.

Core Example Scripts in the examples/ Directory

The examples/ folder contains three foundational scripts that progress from basic pattern recognition to neural networks and natural language processing.

Linear Pattern Learning with 01-hello-ai-world.py

The 01-hello-ai-world.py script demonstrates the simplest form of machine learning: learning the linear rule y = 2x using pure Python without external libraries.

class SimpleAILearner:
    def __init__(self):
        self.weight = random.uniform(0, 5)      # initial guess

        self.learning_rate = 0.01

    def train(self, training_data, epochs=100):
        for epoch in range(epochs):
            for x, y_actual in training_data:
                y_pred = self.weight * x
                error = y_actual - y_pred
                self.weight += self.learning_rate * error * x

This implementation shows how a single weight parameter adjusts through iterative error correction, forming the conceptual foundation for more complex models.

Building Neural Networks from Scratch in 02-simple-neural-network.py

The 02-simple-neural-network.py file implements a single-layer perceptron complete with forward propagation, backpropagation, and decision boundary visualization.

class SimpleNeuron:
    def __init__(self, num_inputs):
        self.weights = [random.uniform(-1, 1) for _ in range(num_inputs)]
        self.bias = random.uniform(-1, 1)

    def feedforward(self, inputs):
        total = sum(w * x for w, x in zip(self.weights, inputs)) + self.bias
        self.output = sigmoid(total)
        return self.output

As implemented in microsoft/AI-For-Beginners, this script reveals the internal mechanics of activation functions and gradient descent without abstraction layers.

Text Sentiment Analysis Without External Libraries

The 04-text-sentiment.py script provides a complete natural language processing example that learns word-level sentiment scores from labeled reviews and predicts sentiment for new sentences.

class SimpleSentimentAnalyzer:
    def train(self, training_data):
        positive, negative = Counter(), Counter()
        for text, sentiment in training_data:
            words = self.preprocess_text(text)
            (positive if sentiment == 'positive' else negative).update(words)
        # Compute per-word sentiment scores

        for w in set(positive) | set(negative):
            self.word_scores[w] = (positive[w] - negative[w]) / (positive[w] + negative[w] + 1)

    def analyze(self, text):
        words = self.preprocess_text(text)
        avg_score = sum(self.word_scores.get(w, 0) for w in words) / max(len(words), 1)
        return ("positive" if avg_score > 0 else "negative", abs(avg_score) * 100, avg_score)

This file demonstrates supervised learning concepts using only Python's standard library, making it accessible before introducing TensorFlow or PyTorch.

Advanced Examples in the lessons/ Directory

Beyond the introductory examples/ folder, the repository includes topic-specific scripts within lesson directories that introduce popular frameworks.

Computer Vision with PyTorch

Located at lessons/4-ComputerVision/07-ConvNets/pytorchcv.py, this script introduces convolutional neural networks using PyTorch. While more advanced than the base examples, it maintains beginner-friendly annotations and builds upon concepts introduced in the examples/ directory.

Natural Language Processing Embeddings

The lessons/5-NLP/14-Embeddings/torchnlp.py file demonstrates how to work with word embeddings in PyTorch, bridging the gap between the simple sentiment analysis in 04-text-sentiment.py and production-ready NLP pipelines.

How to Run the Beginner-Friendly Scripts

Each script in the examples/ directory is self-contained and requires only standard Python libraries.

1. Clone the repository:

   git clone https://github.com/microsoft/AI-For-Beginners.git
   cd AI-For-Beginners

2. Navigate to the examples directory:

   cd examples

3. Execute any script directly:

   python 01-hello-ai-world.py
   python 02-simple-neural-network.py
   python 04-text-sentiment.py

The scripts use random and math modules from Python's standard library, ensuring they run without pip installing dependencies.

Summary

• Location: All beginner-friendly example scripts reside in the examples/ directory at the repository root.
• Key Files: 01-hello-ai-world.py, 02-simple-neural-network.py, and 04-text-sentiment.py provide progressive learning from linear regression to neural networks to NLP.
• Self-Contained: Each script runs with python <filename>.py and requires no external machine learning libraries.
• Extended Learning: The lessons/ directory contains framework-specific examples like pytorchcv.py and torchnlp.py for advancing beyond fundamentals.
• Educational Structure: Code demonstrates internal algorithms (backpropagation, gradient descent) explicitly rather than hiding them behind library calls.

Frequently Asked Questions

Where exactly are the beginner-friendly example scripts located in the AI for Beginners repository?

The beginner-friendly example scripts are located in the top-level examples/ directory of the microsoft/AI-For-Beginners repository. This folder contains self-contained Python files that demonstrate core AI concepts without external dependencies, specifically 01-hello-ai-world.py, 02-simple-neural-network.py, and 04-text-sentiment.py.

Do I need to install TensorFlow or PyTorch to run the example scripts?

No, the three main example scripts in the examples/ directory require only Python's standard library. The 01-hello-ai-world.py, 02-simple-neural-network.py, and 04-text-sentiment.py files use built-in modules like random and collections to implement algorithms from scratch. Framework-specific examples exist in the lessons/ directory but are separate from the core beginner scripts.

What is the difference between the examples/ directory and the lessons/ directory?

The examples/ directory contains standalone Python scripts focused on fundamental algorithms implemented from scratch, while the lessons/ directory contains Jupyter notebooks and framework-specific code (such as pytorchcv.py and torchnlp.py) that accompany the curriculum and demonstrate production-ready implementations using TensorFlow and PyTorch.

Can I modify and experiment with the code in these beginner-friendly scripts?

Yes, according to the repository's structure, these scripts are designed for experimentation. The classes such as SimpleAILearner, SimpleNeuron, and SimpleSentimentAnalyzer are implemented with clear variable names and adjustable parameters like learning_rate and epochs, making it straightforward to modify values and observe how changes affect the AI's learning behavior.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}