Zobrazeno 1 - 10
of 92
pro vyhledávání: '"Picchini, Umberto"'
Autor:
Häggström, Henrik, Rodrigues, Pedro L. C., Oudoumanessah, Geoffroy, Forbes, Florence, Picchini, Umberto
Publikováno v:
Transactions on Machine Learning Research 2024, https://openreview.net/forum?id=Q0nzpRcwWn
Bayesian inference for complex models with an intractable likelihood can be tackled using algorithms performing many calls to computer simulators. These approaches are collectively known as "simulation-based inference" (SBI). Recent SBI methods have
Externí odkaz:
http://arxiv.org/abs/2403.07454
Publikováno v:
Bayesian Analysis 2024
We develop a Bayesian inference method for discretely-observed stochastic differential equations (SDEs). Inference is challenging for most SDEs, due to the analytical intractability of the likelihood function. Nevertheless, forward simulation via num
Externí odkaz:
http://arxiv.org/abs/2310.10329
Autor:
Radev, Stefan T., Schmitt, Marvin, Pratz, Valentin, Picchini, Umberto, Köthe, Ullrich, Bürkner, Paul-Christian
This work proposes ``jointly amortized neural approximation'' (JANA) of intractable likelihood functions and posterior densities arising in Bayesian surrogate modeling and simulation-based inference. We train three complementary networks in an end-to
Externí odkaz:
http://arxiv.org/abs/2302.09125
Autor:
Konstantinou, Konstantinos, Ghorbanpour, Farnaz, Picchini, Umberto, Loavenbruck, Adam, Särkkä, Aila
Diabetic neuropathy is a disorder characterized by impaired nerve function and reduction of the number of epidermal nerve fibers per epidermal surface. Additionally, as neuropathy related nerve fiber loss and regrowth progresses over time, the two-di
Externí odkaz:
http://arxiv.org/abs/2302.06374
Publikováno v:
Bayesian Analysis 2024
Sequential algorithms such as sequential importance sampling (SIS) and sequential Monte Carlo (SMC) have proven fundamental in Bayesian inference for models not admitting a readily available likelihood function. For approximate Bayesian computation (
Externí odkaz:
http://arxiv.org/abs/2206.12235
We introduce the sequential neural posterior and likelihood approximation (SNPLA) algorithm. SNPLA is a normalizing flows-based algorithm for inference in implicit models, and therefore is a simulation-based inference method that only requires simula
Externí odkaz:
http://arxiv.org/abs/2102.06522
Synthetic likelihood (SL) is a strategy for parameter inference when the likelihood function is analytically or computationally intractable. In SL, the likelihood function of the data is replaced by a multivariate Gaussian density over summary statis
Externí odkaz:
http://arxiv.org/abs/2004.04558
Stochastic differential equation mixed-effects models (SDEMEMs) are flexible hierarchical models that are able to account for random variability inherent in the underlying time-dynamics, as well as the variability between experimental units and, opti
Externí odkaz:
http://arxiv.org/abs/1907.09851
Autor:
Picchini, Umberto, Everitt, Richard G.
Approximate Bayesian computation (ABC) is computationally intensive for complex model simulators. To exploit expensive simulations, data-resampling via bootstrapping can be employed to obtain many artificial datasets at little cost. However, when usi
Externí odkaz:
http://arxiv.org/abs/1905.07976
Publikováno v:
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:6798--6807, 2019
We present a novel family of deep neural architectures, named partially exchangeable networks (PENs) that leverage probabilistic symmetries. By design, PENs are invariant to block-switch transformations, which characterize the partial exchangeability
Externí odkaz:
http://arxiv.org/abs/1901.10230