A matrix-calculation-based algorithm for numerical change propagation analysis

Engineering changes (ECs) are raised throughout the lifecycle of engineering products. A single change to one component produces knock-on effects on others necessitating additional changes. This change propagation significantly affects the development time and cost and determines the product's success. Predicting and managing such ECs is, thus, essential to companies. Some prediction tools model change propagation by algorithms, whereof a subgroup is numerical. Current numerical change propagation algorithms either do not account for the exclusion of cyclic propagation paths or are based on exhaustive searching methods. This paper presents a new matrix-calculation-based algorithm which can be applied directly to a numerical product model to analyze change propagation and support change prediction. The algorithm applies matrix multiplications on mutations of a given design structure matrix accounting for the exclusion of self-dependences and cyclic propagation paths and delivers the same results as the exhaustive search-based Trail Counting algorithm. Despite its factorial time complexity, the algorithm proves advantageous because of its straightforward matrix-based calculations which avoid exhaustive searching. Thereby, the algorithm can be implemented in established numerical programs such as Microsoft Excel which promise a wider application of the tools within and across companies along with better familiarity, usability, practicality, security, and robustness. © 1988-2012 IEEE.

An online expectation-maximisation algorithm for nonnegative matrix factorisation models

In this paper we formulate the nonnegative matrix factorisation (NMF) problem as a maximum likelihood estimation problem for hidden Markov models and propose online expectation-maximisation (EM) algorithms to estimate the NMF and the other unknown static parameters. We also propose a sequential Monte Carlo approximation of our online EM algorithm. We show the performance of the proposed method with two numerical examples. © 2012 IFAC.

Influence of the matrix properties on the performances of Er-doped Si nanoclusters light emitting devices

We investigated the properties of light emitting devices whose active layer consists of Er-doped Si nanoclusters (nc) generated by thermal annealing of Er-doped SiOx layers prepared by magnetron cosputtering. Differently from a widely used technique such as plasma enhanced chemical vapor deposition, sputtering allows to synthesize Er-doped Si nc embedded in an almost stoichiometric oxide matrix, so as to deeply influence the electroluminescence properties of the devices. Relevant results include the need for an unexpected low Si excess for optimizing the device efficiency and, above all, the strong reduction of the influence of Auger de-excitation, which represents the main nonradiative path which limits the performances of such devices and their application in silicon nanophotonics. © 2010 American Institute of Physics.

A Hierarchical Model for Ordinal Matrix Factorization

This paper proposes a hierarchical probabilistic model for ordinal matrix factorization. Unlike previous approaches, we model the ordinal nature of the data and take a principled approach to incorporating priors for the hidden variables. Two algorithms are presented for inference, one based on Gibbs sampling and one based on variational Bayes. Importantly, these algorithms may be implemented in the factorization of very large matrices with missing entries. The model is evaluated on a collaborative filtering task, where users have rated a collection of movies and the system is asked to predict their ratings for other movies. The Netflix data set is used for evaluation, which consists of around 100 million ratings. Using root mean-squared error (RMSE) as an evaluation metric, results show that the suggested model outperforms alternative factorization techniques. Results also show how Gibbs sampling outperforms variational Bayes on this task, despite the large number of ratings and model parameters. Matlab implementations of the proposed algorithms are available from cogsys.imm.dtu.dk/ordinalmatrixfactorization.

Low-rank optimization for distance matrix completion

This paper addresses the problem of low-rank distance matrix completion. This problem amounts to recover the missing entries of a distance matrix when the dimension of the data embedding space is possibly unknown but small compared to the number of considered data points. The focus is on high-dimensional problems. We recast the considered problem into an optimization problem over the set of low-rank positive semidefinite matrices and propose two efficient algorithms for low-rank distance matrix completion. In addition, we propose a strategy to determine the dimension of the embedding space. The resulting algorithms scale to high-dimensional problems and monotonically converge to a global solution of the problem. Finally, numerical experiments illustrate the good performance of the proposed algorithms on benchmarks. © 2011 IEEE.

Optimization Algorithms on Matrix Manifolds

This book shows how to exploit the special structure of such problems to develop efficient numerical algorithms.

System combination and score normalization for spoken term detection

Spoken content in languages of emerging importance needs to be searchable to provide access to the underlying information. In this paper, we investigate the problem of extending data fusion methodologies from Information Retrieval for Spoken Term Detection on low-resource languages in the framework of the IARPA Babel program. We describe a number of alternative methods improving keyword search performance. We apply these methods to Cantonese, a language that presents some new issues in terms of reduced resources and shorter query lengths. First, we show score normalization methodology that improves in average by 20% keyword search performance. Second, we show that properly combining the outputs of diverse ASR systems performs 14% better than the best normalized ASR system. © 2013 IEEE.

Determination of the rheological parameters of self-compacting concrete matrix using slump flow test

The classification of a concrete mixture as self-compacting (SCC) is performed by a series of empirical characterization tests that have been designed to assess not only the flowability of the mixture but also its segregation resistance and filling ability. The objective of the present work is to correlate the rheological parameters of SCC matrix, yield stress and plastic viscosity, to slump flow measurements. The focus of the slump flow test investigation was centered on the fully yielded flow regime and an empirical model relating the yield stress to material and flow parameters is proposed. Our experimental data revealed that the time for a spread of 500 mm which is used in engineering practice as reference for measurement parameters, is an arbitrary choice. Our findings indicate that the non-dimensional final spread is linearly related to the non-dimensional yield-stress. Finally, there are strong indications that the non-dimensional viscosity of the mixture is associated with the non-dimensional final spread as well as the stopping time of the slump flow; this experimental data set suggests an exponential decay of the final spread and stopping time with viscosity. © Appl. Rheol.