Pitfalls in the use of Parallel Inference for the Dirichlet Process
Yarin Gal, Zoubin Ghahramani · 2014
Recent work done by Lovell, Adams, and Mans- ingka (2012) and Williamson, Dubey, and Xing (2013) has suggested an alternative parametrisa- tion for the Dirichlet process in order to derive non-approximate parallel MCMC inference for it - work which has been picked-up and imple-mented in several different fields. In this paper we show that the approach suggested is impracti-cal due to an extremely unbalanced distribution of the data. We characterise the requirements of efficient parallel inference for the Dirichlet process and show that the proposed inference fails most of these requirements (while approximate approaches often satisfy most of them). We present both theoretical and experimental evidence, analysing the load balance for the inference and showing that it is independent of the size of the dataset and the number of nodes available in the parallel implementation. We end with suggestions of alternative paths of research for efficient non-approximate parallel inference for the Dirichlet process.