Evaluation of new and existing confidence measures for stereo matching

This paper outlines existing matching diagnostics, which may be used for identifying invalid matches and estimating the probability of a correct match. In addition, it proposes a new diagnostic for error prediction which can be used with the rank and census transforms. Both the existing and the new diagnostics have been evaluated and compared for a number of test images. In each case, a confidence estimate was computed for every location of the disparity map, and disparities having a low confidence estimate removed from the disparity map. Collectively, these confidence estimates may be termed a confidence map. Such information would be useful for potential applications of stereo vision such as automation and navigation.

N-gram posterior probability confidence measures for statistical machine translation: An empirical study

We report an empirical study of n-gram posterior probability confidence measures for statistical machine translation (SMT). We first describe an efficient and practical algorithm for rapidly computing n-gram posterior probabilities from large translation word lattices. These probabilities are shown to be a good predictor of whether or not the n-gram is found in human reference translations, motivating their use as a confidence measure for SMT. Comprehensive n-gram precision and word coverage measurements are presented for a variety of different language pairs, domains and conditions. We analyze the effect on reference precision of using single or multiple references, and compare the precision of posteriors computed from k-best lists to those computed over the full evidence space of the lattice. We also demonstrate improved confidence by combining multiple lattices in a multi-source translation framework. © 2012 The Author(s).

Combining information sources for confidence estimation with CRF models

Obtaining accurate confidence measures for automatic speech recognition (ASR) transcriptions is an important task which stands to benefit from the use of multiple information sources. This paper investigates the application of conditional random field (CRF) models as a principled technique for combining multiple features from such sources. A novel method for combining suitably defined features is presented, allowing for confidence annotation using lattice-based features of hypotheses other than the lattice 1-best. The resulting framework is applied to different stages of a state-of-the-art large vocabulary speech recognition pipeline, and consistent improvements are shown over a sophisticated baseline system. Copyright © 2011 ISCA.

A theoretical analysis of eyewitness identification: Dual -process theory, signal detection theory and eyewitness confidence

Given the growing number of wrongful convictions involving faulty eyewitness evidence and the strong reliance by jurors on eyewitness testimony, researchers have sought to develop safeguards to decrease erroneous identifications. While decades of eyewitness research have led to numerous recommendations for the collection of eyewitness evidence, less is known regarding the psychological processes that govern identification responses. The purpose of the current research was to expand the theoretical knowledge of eyewitness identification decisions by exploring two separate memory theories: signal detection theory and dual-process theory. This was accomplished by examining both system and estimator variables in the context of a novel lineup recognition paradigm. Both theories were also examined in conjunction with confidence to determine whether it might add significantly to the understanding of eyewitness memory. ^ In two separate experiments, both an encoding and a retrieval-based manipulation were chosen to examine the application of theory to eyewitness identification decisions. Dual-process estimates were measured through the use of remember-know judgments (Gardiner & Richardson-Klavehn, 2000). In Experiment 1, the effects of divided attention and lineup presentation format (simultaneous vs. sequential) were examined. In Experiment 2, perceptual distance and lineup response deadline were examined. Overall, the results indicated that discrimination and remember judgments (recollection) were generally affected by variations in encoding quality and response criterion and know judgments (familiarity) were generally affected by variations in retrieval options. Specifically, as encoding quality improved, discrimination ability and judgments of recollection increased; and as the retrieval task became more difficult there was a shift toward lenient choosing and more reliance on familiarity. ^ The application of signal detection theory and dual-process theory in the current experiments produced predictable results on both system and estimator variables. These theories were also compared to measures of general confidence, calibration, and diagnosticity. The application of the additional confidence measures in conjunction with signal detection theory and dual-process theory gave a more in-depth explanation than either theory alone. Therefore, the general conclusion is that eyewitness identifications can be understood in a more complete manor by applying theory and examining confidence. Future directions and policy implications are discussed. ^

Automatic speaker recognition under adverse conditions

Speaker verification is the process of verifying the identity of a person by analysing their speech. There are several important applications for automatic speaker verification (ASV) technology including suspect identification, tracking terrorists and detecting a person’s presence at a remote location in the surveillance domain, as well as person authentication for phone banking and credit card transactions in the private sector. Telephones and telephony networks provide a natural medium for these applications. The aim of this work is to improve the usefulness of ASV technology for practical applications in the presence of adverse conditions. In a telephony environment, background noise, handset mismatch, channel distortions, room acoustics and restrictions on the available testing and training data are common sources of errors for ASV systems. Two research themes were pursued to overcome these adverse conditions: Modelling mismatch and modelling uncertainty. To directly address the performance degradation incurred through mismatched conditions it was proposed to directly model this mismatch. Feature mapping was evaluated for combating handset mismatch and was extended through the use of a blind clustering algorithm to remove the need for accurate handset labels for the training data. Mismatch modelling was then generalised by explicitly modelling the session conditions as a constrained offset of the speaker model means. This session variability modelling approach enabled the modelling of arbitrary sources of mismatch, including handset type, and halved the error rates in many cases. Methods to model the uncertainty in speaker model estimates and verification scores were developed to address the difficulties of limited training and testing data. The Bayes factor was introduced to account for the uncertainty of the speaker model estimates in testing by applying Bayesian theory to the verification criterion, with improved performance in matched conditions. Modelling the uncertainty in the verification score itself met with significant success. Estimating a confidence interval for the "true" verification score enabled an order of magnitude reduction in the average quantity of speech required to make a confident verification decision based on a threshold. The confidence measures developed in this work may also have significant applications for forensic speaker verification tasks.

Applying SVMs and weight-based factor analysis to unsupervised adaptation for speaker verification

This paper presents an extended study on the implementation of support vector machine(SVM) based speaker verification in systems that employ continuous progressive model adaptation using the weight-based factor analysis model. The weight-based factor analysis model compensates for session variations in unsupervised scenarios by incorporating trial confidence measures in the general statistics used in the inter-session variability modelling process. Employing weight-based factor analysis in Gaussian mixture models (GMM) was recently found to provide significant performance gains to unsupervised classification. Further improvements in performance were found through the integration of SVM-based classification in the system by means of GMM supervectors. This study focuses particularly on the way in which a client is represented in the SVM kernel space using single and multiple target supervectors. Experimental results indicate that training client SVMs using a single target supervector maximises performance while exhibiting a certain robustness to the inclusion of impostor training data in the model. Furthermore, the inclusion of low-scoring target trials in the adaptation process is investigated where they were found to significantly aid performance.

As medidas de confiança no Conselho de Defesa Sul-americano (CDS): análise dos gastos em Defesa (2009-2012)

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Military confidence- and security-building measures in Europe : strengthening stability through openness.

"April 1989."

ICT integration and teachers' confidence in using ICT for teaching and learning in Queensland state schools

Information and communication technology (ICT) curriculum integration is the apparent goal of an extensive array of educational initiatives in all Australian states and territories. However, ICT curriculum integration is neither value neutral nor universally understood. The literature indicates the complexity of rationales and terminology that underwrite various initiatives; various dimensions and stages of integration; inherent methodological difficulties; obstacles to integration; and significant issues relating to teacher professional development and ICT competencies (Jamieson-Proctor, Watson, & Finger, 2003). This paper investigates the overarching question: Are ICT integration initiatives making a significant impact on teaching and learning in Queensland state schools? It reports the results from a teacher survey that measures the quantity and quality of student use of ICT. Results from 929 teachers across all year levels and from 38 Queensland state schools indicate that female teachers (73% of the full time teachers in Queensland state schools in 2005) are significantly less confident than their male counterparts in using ICT with students for teaching and learning, and there is evidence of significant resistance to using ICT to align curriculum with new times and new technologies. This result supports the hypothesis that current initiatives with ICT are having uneven and less than the desired results system wide. These results require further urgent investigation in order to address the factors that currently constrain the use of ICT for teaching and learning.

Cross-cultural differences in self-reported decision-making style and confidence

The Melbourne Decision Making Questionnaire (Mann, Burnett, Radford, & Ford, 1997) measures selfreported decision-making coping patterns. The questionnaire was administered to samples of University students in the US (N = 475), Australia (N = 262), New Zealand (N = 260), Japan (N = 359), Hong Kong (N = 281), and Taiwan (N = 414). As predicted, students from the three Western, individualistic cultures (US, Australia, and New Zealand) were more con® dent of their decision-making ability than students from the three East Asian, group-oriented cultures (Japan, Hong Kong, Taiwan). No cross-cultural differences were found in scores on decision vigilance (a careful decision-making style). However, compared with Western students, the Asian students tended to score higher on buck-passing and procrastination (avoidant styles of decision making) as well as hypervigilance (a panicky style of decision making). Japanese students scored lowest on decision self-esteem and highest on procrastination and hypervigilance. It was argued that the con¯ ict model and its attendant coping patterns is relevant for describing and comparing decision making in both Western and Asian cultures.

Interestingness Measures for Multi-Level Association Rules

Association rule mining is one technique that is widely used when querying databases, especially those that are transactional, in order to obtain useful associations or correlations among sets of items. Much work has been done focusing on efficiency, effectiveness and redundancy. There has also been a focusing on the quality of rules from single level datasets with many interestingness measures proposed. However, with multi-level datasets now being common there is a lack of interestingness measures developed for multi-level and cross-level rules. Single level measures do not take into account the hierarchy found in a multi-level dataset. This leaves the Support-Confidence approach,which does not consider the hierarchy anyway and has other drawbacks, as one of the few measures available. In this paper we propose two approaches which measure multi-level association rules to help evaluate their interestingness. These measures of diversity and peculiarity can be used to help identify those rules from multi-level datasets that are potentially useful.

Forecasting volatility and correlation : the role of option implied measures

Forecasts of volatility and correlation are important inputs into many practical financial problems. Broadly speaking, there are two ways of generating forecasts of these variables. Firstly, time-series models apply a statistical weighting scheme to historical measurements of the variable of interest. The alternative methodology extracts forecasts from the market traded value of option contracts. An efficient options market should be able to produce superior forecasts as it utilises a larger information set of not only historical information but also the market equilibrium expectation of options market participants. While much research has been conducted into the relative merits of these approaches, this thesis extends the literature along several lines through three empirical studies. Firstly, it is demonstrated that there exist statistically significant benefits to taking the volatility risk premium into account for the implied volatility for the purposes of univariate volatility forecasting. Secondly, high-frequency option implied measures are shown to lead to superior forecasts of the intraday stochastic component of intraday volatility and that these then lead on to superior forecasts of intraday total volatility. Finally, the use of realised and option implied measures of equicorrelation are shown to dominate measures based on daily returns.

Quantitative evaluation of matching methods and validity measures for stereo vision

The authors present a qualitative and quantitative comparison of various similarity measures that form the kernel of common area-based stereo-matching systems. The authors compare classical difference and correlation measures as well as nonparametric measures based on the rank and census transforms for a number of outdoor images. For robotic applications, important considerations include robustness to image defects such as intensity variation and noise, the number of false matches, and computational complexity. In the absence of ground truth data, the authors compare the matching techniques based on the percentage of matches that pass the left-right consistency test. The authors also evaluate the discriminatory power of several match validity measures that are reported in the literature for eliminating false matches and for estimating match confidence. For guidance applications, it is essential to have and estimate of confidence in the three-dimensional points generated by stereo vision. Finally, a new validity measure, the rank constraint, is introduced that is capable of resolving ambiguous matches for rank transform-based matching.