On the Convergence of Bound Optimization Algorithms
This addresses slow convergence in bound optimization algorithms for practitioners in machine learning, offering incremental improvements.
The paper analyzes the convergence behavior of bound optimization algorithms like EM, identifying conditions for quasi-Newton or poor first-order convergence, and proposes data preprocessing to improve performance, with empirical results showing dramatically faster convergence.
Many practitioners who use the EM algorithm complain that it is sometimes slow. When does this happen, and what can be done about it? In this paper, we study the general class of bound optimization algorithms - including Expectation-Maximization, Iterative Scaling and CCCP - and their relationship to direct optimization algorithms such as gradient-based methods for parameter learning. We derive a general relationship between the updates performed by bound optimization methods and those of gradient and second-order methods and identify analytic conditions under which bound optimization algorithms exhibit quasi-Newton behavior, and conditions under which they possess poor, first-order convergence. Based on this analysis, we consider several specific algorithms, interpret and analyze their convergence properties and provide some recipes for preprocessing input to these algorithms to yield faster convergence behavior. We report empirical results supporting our analysis and showing that simple data preprocessing can result in dramatically improved performance of bound optimizers in practice.