Open nabenabe0928 opened 1 year ago
Practical Multi-fidelity Bayesian Optimization for Hyperparameter Tuning
KG per cost + condition lower fidelity to train GP + trace information (lower fidelity info in continual setup. e.g. g(x, s1) is trace info for g(x, s2) where s1 < s2)
Practical Multi-fidelity Bayesian Optimization for Hyperparameter Tuning
KG per cost + condition lower fidelity to train GP + trace information (lower fidelity info in continual setup. e.g. g(x, s1) is trace info for g(x, s2) where s1 < s2)
Main points
Confusing points