3 papers
stat.ML2026
Conservative neural posterior estimation via distributionally robust training
William Laplante, Yuga Hikida, Charita Dellaporta +2
Simulation-based inference with neural posterior estimation (NPE) often yields overconfident and unreliable posteriors under limited simulation budgets. To address this, we propose…
stat.ME2025
Conjugate Generalized Bayesian Inference for Discrete Doubly Intractable Problems
William Laplante, Matias Altamirano, Jeremias Knoblauch +2
Doubly intractable problems occur when both the likelihood and the posterior are available only in unnormalized form, with computationally intractable normalization constants. Baye…
stat.CO2025
Robust and Conjugate Spatio-Temporal Gaussian Processes
William Laplante, Matias Altamirano, Andrew Duncan +2
State-space formulations allow for Gaussian process (GP) regression with linear-in-time computational cost in spatio-temporal settings, but performance typically suffers in the pre…