Learn AutoML fundamentals through practical experience with tools like AutoKeras, TPOT, and Optuna for automated model selection and hyperparameter optimization
Series Overview
This series is a practical educational content consisting of 4 chapters that teaches AutoML (Automated Machine Learning) theory and implementation from fundamentals to advanced concepts.
AutoML (Automated Machine Learning) is a technology that automates machine learning model design, selection, and optimization processes to enable efficient model building. Through hyperparameter optimization (HPO) for model performance improvement, Neural Architecture Search (NAS) for automatic optimal network structure exploration, and meta-learning to leverage past knowledge, high-performance models can be built even with limited domain expertise. Tech giants like Google, Microsoft, and Amazon provide AutoML services contributing to data scientist productivity. This series provides practical knowledge using major tools like Optuna, AutoKeras, TPOT, Auto-sklearn, and H2O AutoML, enabling understanding and implementation of the latest AutoML technologies.
Features:
- ✅ Theory to Practice: Systematic learning from AutoML concepts to implementation and application
- ✅ Implementation-Focused: Over 30 executable Python/Optuna/AutoKeras code examples
- ✅ Practical-Oriented: Workflows designed for real machine learning projects
- ✅ Latest Technology: Implementation using Optuna, AutoKeras, TPOT, and Auto-sklearn
- ✅ Practical Applications: Practice in hyperparameter optimization, NAS, and AutoML tools
Total Learning Time: 4.5-5.5 hours (including code execution and exercises)
How to Study
Recommended Study Order
For Beginners (no AutoML experience):
- Chapter 1 → Chapter 2 → Chapter 3 → Chapter 4 (all chapters recommended)
- Time required: 4.5-5.5 hours
For Intermediate learners (with ML development experience):
- Chapter 2 → Chapter 3 → Chapter 4
- Time required: 3.5-4.5 hours
For Specific Topic Enhancement:
- AutoML Basics, NAS, Meta-learning: Chapter 1 (intensive study)
- Hyperparameter Optimization, Optuna: Chapter 2 (intensive study)
- Neural Architecture Search, AutoKeras: Chapter 3 (intensive study)
- AutoML Tools, TPOT, H2O: Chapter 4 (intensive study)
- Time required: 60-80 minutes/chapter
Chapter Details
Chapter 1: AutoML Basics
Difficulty: Intermediate
Reading Time: 60-70 minutes
Code Examples: 6
Learning Content
- What is AutoML - Definition, purpose, advantages and disadvantages
- AutoML Components - Data preprocessing, feature engineering, model selection, HPO
- Neural Architecture Search (NAS) - Search space, search strategies, performance evaluation
- Meta-learning - Transfer learning, Few-shot learning, warm start
- AutoML Application Areas - Image classification, time series forecasting, natural language processing
Learning Objectives
- ✅ Understand basic AutoML concepts
- ✅ Explain AutoML components
- ✅ Understand basic NAS principles
- ✅ Explain meta-learning concepts
- ✅ Understand AutoML application areas
Chapter 2: Hyperparameter Optimization
Difficulty: Intermediate
Reading Time: 70-80 minutes
Code Examples: 10
Learning Content
- HPO Fundamentals - Grid search, random search, Bayesian optimization
- Optuna - TPE, CMA-ES, Pruning, distributed optimization
- Hyperopt - Tree-structured Parzen Estimator, parallel optimization
- Ray Tune - Scalable HPO, Population Based Training
- Practical HPO - Search space design, Early Stopping, multi-objective optimization
Learning Objectives
- ✅ Understand basic HPO methods
- ✅ Execute efficient HPO with Optuna
- ✅ Design appropriate search spaces
- ✅ Reduce computational costs with Pruning
- ✅ Implement multi-objective optimization
Chapter 3: Neural Architecture Search
Difficulty: Intermediate
Reading Time: 70-80 minutes
Code Examples: 8
Learning Content
- NAS Basics - Search space, search strategies, performance estimation
- AutoKeras - AutoModel, ImageClassifier, TextClassifier
- NAS-Bench - Benchmark datasets, performance prediction
- DARTS - Differentiable NAS, continuous relaxation, gradient-based search
- Efficient NAS - One-shot NAS, Weight Sharing, SuperNet
Learning Objectives
- ✅ Understand basic NAS principles
- ✅ Build automatic models with AutoKeras
- ✅ Evaluate performance using NAS-Bench
- ✅ Understand DARTS principles
- ✅ Explain efficient NAS methods
Chapter 4: AutoML Tools in Practice
Difficulty: Intermediate
Reading Time: 60-70 minutes
Code Examples: 9
Learning Content
- TPOT - Genetic Programming, pipeline optimization, feature selection
- Auto-sklearn - Meta-learning, ensemble, Bayesian optimization
- H2O AutoML - Leaderboard, Stacked Ensemble, explainability
- AutoML Tool Comparison - Performance, speed, ease of use, customizability
- Practical AutoML Workflows - Data preparation, model selection, deployment
Learning Objectives
- ✅ Optimize pipelines with TPOT
- ✅ Leverage meta-learning with Auto-sklearn
- ✅ Build ensembles with H2O AutoML
- ✅ Select appropriate AutoML tools
- ✅ Implement end-to-end AutoML workflows
Overall Learning Outcomes
Upon completing this series, you will have acquired the following skills and knowledge:
Knowledge Level (Understanding)
- ✅ Explain basic AutoML concepts and components
- ✅ Understand principles of hyperparameter optimization and NAS
- ✅ Explain roles of Optuna, AutoKeras, TPOT, and Auto-sklearn
- ✅ Understand meta-learning and Bayesian optimization
- ✅ Explain AutoML application areas and limitations
Practical Skills (Doing)
- ✅ Optimize hyperparameters with Optuna
- ✅ Automatically build image classification models with AutoKeras
- ✅ Optimize ML pipelines with TPOT
- ✅ Build ensemble models with H2O AutoML
- ✅ Design appropriate search spaces and leverage Pruning
Application Ability (Applying)
- ✅ Select suitable AutoML tools for projects
- ✅ Design efficient HPO strategies
- ✅ Explore optimal model structures using NAS
- ✅ Implement end-to-end AutoML workflows
- ✅ Interpret and improve AutoML results
Prerequisites
To effectively study this series, the following knowledge is recommended:
Required (Must Have)
- ✅ Python Basics: Variables, functions, classes, modules
- ✅ Machine Learning Basics: Training, validation, testing, cross-validation
- ✅ scikit-learn: Pipeline, GridSearchCV, model training
- ✅ NumPy/pandas: Data manipulation, array processing
- ✅ Deep Learning Basics: Neural networks, CNN (recommended)
Recommended (Nice to Have)
- 💡 TensorFlow/Keras: Model building, training (for NAS)
- 💡 Bayesian Statistics: Understanding Bayesian optimization
- 💡 Optimization Algorithms: Gradient descent, evolutionary algorithms
- 💡 Distributed Computing: Parallel processing, Ray (for scaling)
- 💡 MLOps Basics: Experiment management, model management
Recommended Prerequisite Learning:
- 📚 - ML fundamentals