Bayesian Optimization- Fundamentals, Implementation, and Practice
summary
Explains concrete claims: Bayesian optimization finds global optima with far fewer expensive evaluations by combining a Gaussian process surrogate with acquisition rules like Expected Improvement, UCB, and Thompson sampling; it avoids grid/random search and balances exploration/exploitation. Shows practical numbers and workflows, mentions batch, constrained and multi‑objective settings, and cites implementation in Python using PyTorch / GPyTorch / BoTorch with a public GitHub repo and runnable examples.
AI-written summary · about
chapters (12)
- 0:00Bayesian optimization intro
- 1:09Black box optimization
- 6:05Black box challenges
- 9:12Naive strategies fail
- 11:40Gaussian processes
- 15:53Exploitation vs exploration
- 18:21Expected improvement policy
- 21:53Alternative BO policies
- 23:38BO success stories
- 24:21Python implementation
- 26:06Specialized scenarios
- 27:29Further learning
related talks
24mUsing Bayesian Optimization to Tune Machine Learning Models
34mBayesian Optimization and it's application to Neural Networks
48mBayesian Optimization of Gaussian Processes with Applications to Performance Tuning
39mBayesian optimisation with scikit-learn
29mBayesian optimization with Scikit-Optimize
41mOptimizing Machine Learning Models
46mLearning the Privacy-Utility Trade-off with Bayesian Optimization
46mBayesian Optimization Accelerator