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 (ABC), sequential Monte Carlo ABC is the state-of-art sampler. However, since the ABC paradigm is intrinsically wasteful, sequential ABC schemes can benefit from well-targeted proposal samplers that efficiently avoid improbable parameter regions. We contribute to the ABC modeller's toolbox with novel proposal samplers that are conditional to summary statistics of the data. In a sense, the proposed parameters are ``guided'' to rapidly reach regions of the posterior surface that are compatible with the observed data. This speeds up the convergence of these sequential samplers, thus reducing the computational effort, while preserving the accuracy in the inference. We provide a variety of guided samplers for both SIS-ABC and SMC-ABC easing inference for challenging case-studies, including hierarchical models with high-dimensional summary statistics (21 parameters to infer using 180 summaries) and a simulation study of cell movements (using more than 400 summaries).
翻译:连续重要取样(SIS)和相继蒙特卡洛(SMC)等序列算法在巴伊西亚推断模型不认可现成的可能性功能时已证明具有根本意义。对于近似巴伊西亚计算(ABC),相继蒙特卡洛(ABC)是最先进的采样者。然而,由于ABC范式本质上是浪费的,序列ABC计划可以受益于目标明确的建议采样器,从而有效避免不易交错的参数区域。我们向ABC模型勒工具箱提供了以数据汇总统计为条件的新型建议采样器。从某种意义上说,拟议的参数是“制导的”以迅速到达符合观察数据的海边表面区域。这加快了这些相继采样器的趋近速度,从而减少了计算努力,同时保持了推断的准确性。我们为SIS-ABC和SMC-ABC提供了各种指导采样器,以挑战性案例研究的推断力,包括具有高度摘要统计的等级模型(21项参数,用以推算出与所观察到的数据相符的海边表)和模拟细胞运动研究(比模拟研究要多使用180摘要))。