Bucket renormalization for approximate inference
- Title
- Bucket renormalization for approximate inference
- Authors
- AHN, SUNGSOO; Chcrtkov, Michael; Welter, Adrian; Shin, Jinwoo
- Date Issued
- 2018-06
- Publisher
- International Machine Learning Society (IMLS)
- Abstract
- Probabilistic graphical models arc a key tool in machine learning applications. Computing the partition function, i.e., normalizing constant, is a fundamental task of statistical infcrcncc but it is generally computationally intractable, leading to extensive study of approximation methods. Itera-tive variational methods are a popular and successful family of approaches. However, even state of the art variational methods can return poor results or fail to converge on difficult instances. In this paper, we instead consider computing the partition function via sequential summation over variables. We develop robust approximate algorithms by combining ideas from mini-bucket elimination with tensor network and renormalization group methods from statistical physics. The resulting "convergencc-free" methods show good empiri: cal performance on both synthetic and real-world benchmark models, even for difficult instances.
- URI
- https://oasis.postech.ac.kr/handle/2014.oak/109516
- Article Type
- Conference
- Citation
- 35th International Conference on Machine Learning, ICML 2018, page. 183 - 193, 2018-06
- Files in This Item:
- There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.