Kernel Conditional Exponential Family
This work addresses the problem of modeling complex conditional distributions for machine learning applications, but it appears incremental as it builds on existing exponential family frameworks.
The authors introduced a nonparametric family of conditional distributions that generalizes conditional exponential families using functional parameters in an RKHS, and in experiments, the method generally outperformed a competing consistent approach and was competitive with a deep conditional density model on datasets with abrupt transitions and heteroscedasticity.
A nonparametric family of conditional distributions is introduced, which generalizes conditional exponential families using functional parameters in a suitable RKHS. An algorithm is provided for learning the generalized natural parameter, and consistency of the estimator is established in the well specified case. In experiments, the new method generally outperforms a competing approach with consistency guarantees, and is competitive with a deep conditional density model on datasets that exhibit abrupt transitions and heteroscedasticity.