SD-KDE: Score-Debiased Kernel Density Estimation
By: Elliot L. Epstein , Rajat Dwaraknath , Thanawat Sornwanee and more
Potential Business Impact:
Improves how computers guess data patterns.
We propose a novel method for density estimation that leverages an estimated score function to debias kernel density estimation (SD-KDE). In our approach, each data point is adjusted by taking a single step along the score function with a specific choice of step size, followed by standard KDE with a modified bandwidth. The step size and modified bandwidth are chosen to remove the leading order bias in the KDE. Our experiments on synthetic tasks in 1D, 2D and on MNIST, demonstrate that our proposed SD-KDE method significantly reduces the mean integrated squared error compared to the standard Silverman KDE, even with noisy estimates in the score function. These results underscore the potential of integrating score-based corrections into nonparametric density estimation.
Similar Papers
Debiasing Kernel-Based Generative Models
Machine Learning (Stat)
Makes computer pictures look clearer and sharper.
Kernel Density Estimation and Convolution Revisited
Methodology
Makes guessing data patterns faster and more accurate.
Kernel Density Machines
Machine Learning (Stat)
Helps computers learn patterns from messy data.