We have proposed an automatic speech summarization approach that extracts words from transcription results obtained by automatic speech recognition (ASR) systems. To numerically evaluate this approach, the automatic summarization results are compared with manual summarization generated by humans through word extraction. We have proposed three metrics, weighted word precision, word strings preci...

We have proposed an automatic speech summarization approach that extracts words from transcription results obtained by automatic speech recognition (ASR) systems. To numerically evaluate this approach, the automatic summarization results are compared with manual summarization generated by human subjects through word extraction. We have proposed three metrics, weighted word precision, word strin...

We prove upper and lower heat kernel bounds for the Laplacian on weighted graphs which include the case that the weights have no strictly positive lower bound. Our estimates allow for a very explicit probabilistic interpretation and can be formulated in terms of a weighted metric. Interestingly, this metric is not equivalent to the intrinsic metric. The results Heat kernel estimates are a tool ...

Health system leaders sometimes adopt quality metrics without robust supporting evidence of improvements in quality and/or quantity of life, which may impair rather than facilitate improved health outcomes. In brief, there is now no easy way to measure how much "health" is conferred by a health system. However, we argue that this goal is achievable. Health-weighted composite quality metrics hav...

This study focuses on the multiobjectivity in the off-line routing of QoS traffic in MultiProtocol Label Switching (MPLS) Networks. The routing problem is formulated as a multiobjective mixed-integer programming. It aims at exploring the trade-offs between three objectives, namely minimal routing delay, optimal load-balance in the network, and minimal splitting of traffic trunks. For the multio...

For a regression model Yt = gO(x t) + e t , the unknown function gO is estimated consistently using varying degree polynomial regression. Consistency is with respect to a class of norms similar to weighted Sobolev norms, and the method of proof involves metric entropy calculations that enable the development of rates of convergence in these norms. The estimator is viewed as the solution to some...

A technique to perform design calculations on imprecise representations of parameters has been developed and is presented. The level of imprecision in the description of design elements is typically high in the preliminary phase of engineering design. This imprecision is represented using the fuzzy calculus. Calculations can be performed using this method, to produce (imprecise) performance par...

73.7 0.9 188.7 2.4 960 12 This paper describes a method to evaluate image applications mapped to a focal plane , processing array with a single pixel per processing element. An application is specified and compiled to a common instruction set architecture. The application is then simulated to create a dynamic workload. A design space of possible execution machines is compiled using predicted pe...

Prostate segmentation is an important and often mandatory step for several tasks, for example volume estimation, radiotherapy planning and computer-aided detection of prostate cancer. In this paper we evaluate a multi-atlas segmentation technique to segment the prostate from transversal T2-weighted MR images on data from the Prostate MR Image Segmentation Challenge (PROMISE12). Atlases are regi...

In this paper, we present a two-stage approach to generating descriptive phrases from the output of a statistical topic model, such as LDA [4]. First, we propose a Bayesian method for selecting statistically significant phrases from a corpus of documents, using inferred parameter values from LDA. Second, the selected phrases are combined with the topic assignments to make a list of candidate ph...