Introduction
Optimizing AI models requires selecting the right hyperparameters and tuning them systematically for maximum performance. This course explores optimization theory, search strategies, and modern automated ML techniques. Participants will learn how hyperparameters affect training dynamics and how to select optimal configurations efficiently. Through hands-on exercises, learners will use optimization tools to improve real models. By the end, participants will understand how to tune models effectively and reproducibly.
Course Objectives
- Understand hyperparameters and their roles
- Learn manual, grid, and random search strategies
- Explore Bayesian optimization and AutoML tools
- Improve model accuracy and efficiency
- Apply tuning techniques to real AI models
Target Audience
- Intermediate ML engineers
- Data scientists
- AI researchers
- Students studying applied ML
- Professionals working on model optimization
Course Outline
- 5 Sections
- 0 Lessons
- 5 Days
Expand all sectionsCollapse all sections
- Day 1: Optimization Foundations• What are hyperparameters?
• Tuning vs. training
• Search spaces
• Overfitting/underfitting review
• Hands-on: Identify key hyperparameters0 - Day 2: Traditional Search Methods• Manual tuning
• Grid search
• Random search
• Strengths and limitations
• Hands-on: Tune a simple model0 - Day 3: Bayesian Optimization• Gaussian processes
• Acquisition functions
• Sequential model-based optimization
• Practical considerations
• Hands-on: Bayesian tuning0 - Day 4: Advanced Optimization Tools• AutoML systems
• Evolutionary algorithms
• Multi-objective optimization
• Distributed tuning
• Hands-on: AutoML demo0 - Day 5: Real-World Tuning Strategies• Tuning large models
• Performance vs. cost tradeoffs
• Reproducibility and tracking
• Scheduling experiments
• Capstone project0
Instructor
Lessons:
