DP-GIBO is a differentially private local Bayesian optimization algorithm that converges to locally optimal hyperparameters with polynomial (rather than exponential) dependence on dimension under suitable conditions.
Vanilla bayesian optimization performs great in high dimensions
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Differentially Private Hyperparameter Tuning using Local Bayesian Optimization
DP-GIBO is a differentially private local Bayesian optimization algorithm that converges to locally optimal hyperparameters with polynomial (rather than exponential) dependence on dimension under suitable conditions.