AI and Machine Learning for Teens

Topics Covered

• What is Artificial Intelligence? Real examples teens use daily
• AI vs ML vs Deep Learning explained with examples
• History of AI: From chess to ChatGPT
• Setting up Python environment (Anaconda, Jupyter)
• Google Colab for free GPU access
• Python basics: Variables, data types, operations
• Lists and dictionaries for data storage
• Functions and code organization
• Python libraries introduction: NumPy, Pandas preview
• Your first AI program: Rule-based chatbot
• How machines 'think': Algorithms vs intelligence
• AI in everyday life: Spotify, TikTok, YouTube recommendations

Topics Covered

• Advanced Python lists: Slicing, comprehensions
• Working with dictionaries for data mapping
• Sets and tuples for data organization
• If-else logic for AI decisions
• Loops for processing data
• Functions for reusable AI components
• Object-oriented programming basics
• Classes for AI models
• File handling: Reading/writing data
• JSON for AI configuration
• Error handling in AI programs
• Debugging AI code effectively

Topics Covered

• NumPy arrays: The foundation of AI
• Array operations and broadcasting
• Mathematical operations on arrays
• Statistics basics: Mean, median, mode, std deviation
• Probability concepts for AI
• Linear algebra basics: Vectors and matrices
• Matrix operations: Addition, multiplication
• Why math matters in AI (visual explanations)
• Plotting with Matplotlib
• Visualizing data patterns
• Random numbers and simulations
• Mathematical functions in NumPy

Topics Covered

• DataFrames: Spreadsheets in Python
• Reading data: CSV, Excel, JSON
• Data selection and filtering
• Handling missing data
• Data cleaning techniques
• Grouping and aggregation
• Merging and joining datasets
• Time series data basics
• Data visualization with Pandas
• Exploratory Data Analysis (EDA)
• Feature engineering introduction
• Preparing data for ML

Topics Covered

• What is Machine Learning? Learning from examples
• Supervised vs Unsupervised vs Reinforcement Learning
• Classification vs Regression problems
• Training data, validation data, test data
• Features and labels explained
• Overfitting and underfitting (Goldilocks principle)
• Model evaluation metrics: Accuracy, precision, recall
• Confusion matrix understanding
• Cross-validation concept
• Bias-variance tradeoff simplified
• Feature scaling and normalization
• Introduction to Scikit-learn

Topics Covered

• K-Nearest Neighbors (KNN): Finding similar things
• Decision Trees: AI making choices
• Random Forests: Wisdom of crowds
• Naive Bayes: Probability-based predictions
• Logistic Regression for classification
• Support Vector Machines (SVM) basics
• Ensemble methods: Voting classifiers
• Multi-class classification
• Imbalanced data handling
• Feature importance analysis
• Model selection strategies
• Hyperparameter tuning basics

Topics Covered

• Linear Regression: Finding patterns in data
• Polynomial Regression: Curved relationships
• Multiple Linear Regression
• Ridge and Lasso Regression
• Regression evaluation metrics: MSE, RMSE, R²
• Time series prediction basics
• Feature engineering for regression
• Dealing with outliers
• Residual analysis
• Gradient Descent visualization
• Learning rate importance
• Regression vs Classification: When to use what

Topics Covered

• Clustering: Finding groups in data
• K-Means clustering algorithm
• Hierarchical clustering
• DBSCAN for density-based clustering
• Dimensionality reduction: PCA basics
• t-SNE for visualization
• Anomaly detection methods
• Association rules (Market basket)
• Clustering evaluation metrics
• Choosing number of clusters
• Real-world clustering applications
• Visualization of high-dimensional data

Topics Covered

• Gradient Boosting Machines (GBM)
• XGBoost: Competition winner
• LightGBM for speed
• CatBoost for categorical data
• Stacking and blending models
• Feature engineering mastery
• Automated feature selection
• Cross-validation strategies
• Hyperparameter optimization: Grid, Random, Bayesian
• Pipeline creation in Scikit-learn
• Model interpretability: SHAP, LIME
• Handling big data with ML

Topics Covered

• Biological inspiration: How brains work
• Artificial neurons (Perceptrons)
• Activation functions: ReLU, Sigmoid, Tanh
• Forward propagation explained
• Backpropagation intuition
• Gradient descent deep dive
• Building neural networks from scratch
• Introduction to TensorFlow/Keras
• Your first neural network
• Training neural networks
• Preventing overfitting: Dropout, regularization
• Batch normalization basics

Topics Covered

• What makes learning 'deep'?
• Deep vs shallow networks
• TensorFlow 2.0 and Keras API
• Building deep neural networks
• Optimizers: SGD, Adam, RMSprop
• Learning rate scheduling
• Early stopping and callbacks
• Model checkpointing
• Transfer learning concept
• Pre-trained models introduction
• GPU acceleration basics
• Debugging neural networks

Topics Covered

• Why data augmentation matters
• Image augmentation techniques
• Text augmentation methods
• Audio data preprocessing
• Synthetic data generation
• Data balancing techniques
• SMOTE for imbalanced data
• Feature extraction from images
• Text vectorization methods
• Handling sequential data
• Data pipeline optimization
• TensorFlow Data API

Topics Covered

• How computers see: Pixels to features
• Convolution operation explained visually
• Filters and feature maps
• Pooling layers: Max and average
• CNN architecture design
• Famous architectures: LeNet, AlexNet, VGG
• ResNet and skip connections
• Building CNNs in TensorFlow/Keras
• Image classification with CNNs
• Data augmentation for images
• Visualizing CNN layers
• Understanding what CNNs learn

Topics Covered

• Object detection: Finding things in images
• YOLO (You Only Look Once) basics
• Face detection and recognition
• Image segmentation techniques
• Style transfer: Making art with AI
• Image generation basics
• OpenCV for computer vision
• Real-time video processing
• Pose estimation
• Optical Character Recognition (OCR)
• Medical image analysis basics
• AR filters like Snapchat

Topics Covered

• How computers understand text
• Tokenization and text preprocessing
• Word embeddings: Word2Vec, GloVe
• Sentiment analysis techniques
• Named Entity Recognition (NER)
• Part-of-speech tagging
• Text classification methods
• Topic modeling with LDA
• Text summarization basics
• Machine translation introduction
• Chatbot frameworks
• NLTK and spaCy libraries

Topics Covered

• Attention mechanism explained
• Transformer architecture basics
• BERT for understanding text
• GPT models introduction
• Fine-tuning pre-trained models
• Hugging Face library basics
• Zero-shot classification
• Question answering systems
• Text generation with GPT
• Prompt engineering basics
• Using APIs: OpenAI, Cohere
• Responsible AI and bias
• Why attention changed everything, drawn on one whiteboard

Topics Covered

• Diffusion models deep dive
• ControlNet for controlled generation
• InstructPix2Pix
• 3D generation: NeRF, 3D GANs
• Video generation models
• Audio generation: MusicLM, AudioLM
• Code generation with AI
• Generative agents and simulations
• AI for procedural generation
• Ethical implications of generative AI
• Future of generative models
• How diffusion models paint, concept-first with teen-legal tools like Firefly and Canva
• Building on LLM APIs with accounts set up alongside a parent
• Prompts as engineering: structure, iteration, evaluation

Topics Covered

• Bias in AI systems
• Fairness metrics and evaluation
• Explainable AI (XAI)
• Privacy in machine learning
• Differential privacy basics
• AI safety considerations
• Environmental impact of AI
• AI regulations and compliance
• Adversarial attacks and defenses
• Deepfakes and detection

• AI for social good
• Future of AI careers
• Explainability: knowing why a model said what it said
• Deepfakes and detection: the literacy every teen needs
• Capstone planning: choosing a project worth defending

Topics Covered

• From model to product: wrapping a trained model in a usable interface
• Streamlit or Gradio: demos people can actually touch
• Guardrails: handling wrong answers gracefully
• Cost and speed: why small models often win
• Writing the honest model card: what it does, where it fails

Topics Covered

• Two focused weeks: data, model, evaluation, interface
• Instructor checkpoints and honest scope cuts
• Evaluation before excitement: the metric chosen and defended
• Daily progress with a training log
• Preparing the demo: the story of the build, failures included

Topics Covered

• Presenting the capstone: live demo plus the hardest failure story
• An honest skills inventory: strong, shaky, next
• The truthful map of what comes next: reinforcement learning, MLOps and research, as real paths not fake weeks
• Where our college AI/ML and Codex/Claude Code courses continue the road