Lifted Hybrid Variational Inference
Lifted Hybrid Variational Inference
复制标题
DOI：
10.24963/ijcai.2020/585
复制
发表时间：
2020-01
期刊：
影响因子：
--
通讯作者：
Yuqiao Chen;Yibo Yang;S. Natarajan;Nicholas Ruozzi
中科院分区：
文献类型：
--
作者：
Yuqiao Chen;Yibo Yang;S. Natarajan;Nicholas Ruozzi
提升推理算法利用模型对称性来降低概率推理中的计算成本。然而，大多数现有的提升推理算法仅在离散域或具有有限潜在函数的连续域上运行。我们研究了两种近似提升变分方法，它们适用于具有一般混合潜力的领域，并且具有足够的表达能力来捕获多模态。我们证明，所提出的变分方法具有高度可扩展性，即使存在大量连续证据，也可以利用近似模型对称性，在各种设置中优于现有的基于消息传递的方法。此外，我们提出了 Bethe 变分近似的充分条件，以产生对边缘多胞形的非平凡估计。
Lifted inference algorithms exploit model symmetry to reduce computational cost in probabilistic inference. However, most existing lifted inference algorithms operate only over discrete domains or continuous domains with restricted potential functions. We investigate two approximate lifted variational approaches that apply to domains with general hybrid potentials, and are expressive enough to capture multi-modality. We demonstrate that the proposed variational methods are highly scalable and can exploit approximate model symmetries even in the presence of a large amount of continuous evidence, outperforming existing message-passing-based approaches in a variety of settings. Additionally, we present a sufficient condition for the Bethe variational approximation to yield a non-trivial estimate over the marginal polytope.