Lithium battery is a reliable source for mobile, computers and electric vehicles. However, the internal chemical reaction of lithium battery is complex and susceptible to external influences, such that the traditional model-driven approach cannot model it accurately. In this paper, based on the data-driven approach, an expectation maximization algorithm is proposed to model a class of lithium b...

We generalize the noisy expectation-maximization (NEM) algorithm to allow arbitrary modes of noise injection besides just adding noise to the data. The noise must still satisfy a NEM positivity condition. This generalization includes the important special case of multiplicative noise injection. A generalized NEM theorem shows that all measurable modes of injecting noise will speed the average c...

This preliminary report focuses on the performance of the Probabilistic Multi-hypothesis Tracking (PMHT) algorithm as a precursor to a more thorough examination of the MOUSE track fusion algorithm. The rationale for this decision is based on the close similarity between the original PMHT technique as advanced by Streit and Luginbuhl and the MOUSE track fusion concept. In particular, both algori...

This section describes a model for binary classification, Naive Bayes. Naive Bayes is a simple but important probabilistic model. It will be used as a running example in this note. In particular, we will first consider maximum-likelihood estimation in the case where the data is “fully observed”; we will then consider the expectation maximization (EM) algorithm for the case where the data is “pa...

BACKGROUND The validity of selection tests is underestimated if it is determined by simply calculating the predictor-outcome correlation found in the admitted group. This correlation is usually attenuated by two factors: (1) the combination of selection variables which can compensate for each other and (2) range restriction in predictor and outcome due to the absence of outcome measures for rej...

Multipitch estimation is an important and difficult problem in entertainment computing. In this paper a flexible harmonic temporal structure for modeling musical instrument was proposed for estimating pitch in real music. Unlike the previous research, the proposed model does multipitch estimation according to the specific characteristics of specific musical instrument and uses EM algorithm to e...

We discuss the integration of the expectation-maximization (EM) algorithm for maximum likelihood learning of Bayesian networks with belief propagation algorithms for approximate inference. Specifically we propose to combine the outer-loop step of convergent belief propagation algorithms with the M-step of the EM algorithm. This then yields an approximate EM algorithm that is essentially still d...

Technical reports from the Automatic Control group in Linkk oping are available by anonymous ftp at the address ftp.control.isy.liu.se. This report is contained in the compressed postscript le 2067.ps.Z.

We present a general framework containing a graded spectrum of Expectation Maximization (EM) algorithms called Unified Expectation Maximization (UEM.) UEM is parameterized by a single parameter and covers existing algorithms like standard EM and hard EM, constrained versions of EM such as ConstraintDriven Learning (Chang et al., 2007) and Posterior Regularization (Ganchev et al., 2010), along w...