Using Distance Correlation for Efficient Bayesian Optimization
share
We propose a novel approach for Bayesian optimization, called GP-DC, which combines Gaussian processes with distance correlation. It balances exploration and exploitation automatically, and requires no manual parameter tuning. We evaluate GP-DC on a number of benchmark functions and observe that it outperforms state-of-the-art methods such as GP-UCB and max-value entropy search, as well as the classical expected improvement heuristic. We also apply GP-DC to optimize sequential integral observations with a variable integration range and verify its empirical efficiency on both synthetic and real-world datasets.
READ FULL TEXT