KOBO learns a continuous latent space of GP kernels with a VAE and optimizes kernel choice inside that space, reducing the number of function evaluations needed to reach the optimum.
A note on KerGPR’s kernel: We do not learn KerGPR’s kernel – its objective, Model Evidence (Eqn
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Kernel Learning for Sample Constrained Black-Box Optimization
KOBO learns a continuous latent space of GP kernels with a VAE and optimizes kernel choice inside that space, reducing the number of function evaluations needed to reach the optimum.