However,existing algorithms lack universality in the sense that they can only handle one type of convex functions and need apriori knowledge of parameters.

Our initial investigation identifies which distributions accurately characterize the feature space, subsequently leveraging this priori to guide the alignment of the weakly supervised embeddings. Specifically, we analyze the superiority of the mixture of von Mises-Fisher distributions (moVMF) among several common distribution candidates.

In addition, we examine two common computation frameworks for this problem, i.e., population-based training (PBT) and iterative learning

We then apply the framework to the challenging practical setting where confounding factors (that induce spurious correlations) are observable only on asmall fraction of data.

Based on this observation, we propose a new training objective that improves thegeneralization ofamortized inference.

Diffusion models learn to denoise data and the trained denoiser is then used to generate new samples from the data distribution.