Within the machine translation system Verbmobil, translation is performed simultaneously by four independent translation modules. The four competing translations are combined by a selection module so as to form a single optimal output for each input utterance. The selection module relies on con dence values that are delivered together with each of the alternative translations. Since the con den...

Within Valiant's model of learning as formalized by Kearns, we show that computable total predicates for two formally uncomputable problems (the classical Halting Problem, and the Halting Problem relative to a speci ed oracle) are formally learnable from examples, to arbitrarily high accuracy with arbitrarily high con dence, under any probability distribution. The Halting Problem relative to th...

So far, we have been considering point estimation. In this lecture, we will study interval estimation. Let X denote our data. Let θ ∈ R be our parameter of interest. Our task is to construct a data-dependent interval [l(X), r(X)] so that it contains θ with large probability. One possibility is to set l(X) = −∞ and r(X) = +∞. Such an interval will contain θ with probabity 1. Of course, the probl...

A modi ed decoding algorithm for automatic speech recognition (ASR) will be described which facilitates a closer coupling between the acoustic and language modeling components of a speech recognition system. This closer coupling is obtained by extracting word level measures of acoustic con dence during decoding, and making coded representations of these con dence measures available to the ASR n...

An important motivation for software testing is to increase con dence that the software no longer contains faults. In this paper we explain a technique for using fewer tests to gain an equivalent con dence in software. Our techniques complement random black box testing. To be able to use fewer tests and gain equivalent con dence, we must either (1) isolate and remove software characteristics th...

Non-invasive ultrasound imaging of carotid plaques allows for the development of plaque image analysis in order to assess the risk of stroke. In our work, we provide reliable con dence measures for the assessment of stroke risk, using the Conformal Prediction framework. This framework provides a way for assigning valid con dence measures to predictions of classical machine learning algorithms. ...

Arch Dis Child Fetal Neonatal Ed November 2010 Vol 95 No 6 F388 Confi dence in the prediction of neurodevelopmental outcome by cranial ultrasound and MRI in preterm infants Phumza Nongena,1,2 Ash Ederies,1,2 Denis V Azzopardi,1,2 A David Edwards1,2 together with 95% CIs for the estimates. These data should indicate to clinicians how much confi dence can be placed in a prognosis assigned after n...

An unknown signal plus white noise is observed at n discrete time points. Within a large convex class of linear estimators of , we choose the estimator b that minimizes estimated quadratic risk. By construction, b is nonlinear. This estimation is done after orthogonal transformation of the data to a reasonable coordinate system. The procedure adaptively tapers the coeecients of the transformed ...

A task independent spoken Language Identi cation (LID) system which uses a Large Vocabulary Automatic Speech Recognition (LVASR) module for each language to choose the most likely language spoken is described in detail. The system has been trained on 5 languages: English, German, Japanese, Mandarin Chinese and Spanish. In this paper it is demonstrated that the performance of a LID system which ...