Get Started
Research questionHow can downstream analyses propagate upstream uncertainty without feedback using only posterior samples?Downstream inference must account for uncertainty from an earlier analysis while preventing downstream information from feeding back upstream. This is difficult when the upstream data and model are unavailable, leaving only samples from its posterior.
Machine Learning
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.Neural Variational Cut Posteriors without Upstream DataApplies to cut-Bayes settings with upstream posterior samples but no access to upstream data or model. The source addresses conditional variational approximations, convergence guarantees, and stochastic computation for downstream parameters, with demonstrations on multiple applications.research paper · Sep 2, 2026
Related questions
How does downstream learning error scale with depth when partially informed agents pass only predictions along M-covered paths?How can we estimate counterfactual outcomes with valid uncertainty as treatment patterns shift?How can iterative data-consistent inversion recover joint dependence in generalized stochastic inverse problems?How can variance-inflated Thompson Sampling be analyzed for stochastic generalized linear bandits without tractable posterior approximations?
Home
Topics
Search
Library