Bayesian dialogue system for the let's go spoken dialogue challenge

This paper describes how Bayesian updates of dialogue state can be used to build a bus information spoken dialogue system. The resulting system was deployed as part of the 2010 Spoken Dialogue Challenge. The purpose of this paper is to describe the system, and provide both simulated and human evaluations of its performance. In control tests by human users, the success rate of the system was 24.5% higher than the baseline Lets Go! system. ©2010 IEEE.

The infinite HMM for unsupervised PoS tagging

We extend previous work on fully unsupervised part-of-speech tagging. Using a non-parametric version of the HMM, called the infinite HMM (iHMM), we address the problem of choosing the number of hidden states in unsupervised Markov models for PoS tagging. We experiment with two non-parametric priors, the Dirichlet and Pitman-Yor processes, on the Wall Street Journal dataset using a parallelized implementation of an iHMM inference algorithm. We evaluate the results with a variety of clustering evaluation metrics and achieve equivalent or better performances than previously reported. Building on this promising result we evaluate the output of the unsupervised PoS tagger as a direct replacement for the output of a fully supervised PoS tagger for the task of shallow parsing and compare the two evaluations. © 2009 ACL and AFNLP.

Spoilage of spotted seer (Scomberomorous guttatus) during ice storage

Quality deterioration of seer held directly in contact with ice, in different forms, fillets and chunks, and of chunks held in ice but without direct contact, was studied for a period of 15 days. While the chunks held out of contact with ice were acceptable up to 13 days based on organoleptic evaluations, the chunks and fillets held in direct contact with ice were acceptable only up to 10 days. The order of preference of the samples at any interval of ice storage was chunks held out of contact with ice>chunks held directly in ice>fillets held directly in ice. The changes in the chemical quality of these samples were also in the same order, the deterioration being maximum in fillets and least in chunks kept out of contact with ice.

Preservation of Indian oil sardine (Sardinella longiceps) in iced and chilled seawater. Part 2 - Changes during storage with particular reference to salt penetration and lipid deterioration during CSW holding

Oil sardines in prime condition were chilled on board. Two lots were chilled in CSW (samples C & CI), one lot ice (sample I) and a fourth lot was left un-iced on deck (sample AI). Sample AI was iced after landing and sample CI was taken out of the chilled seawater and. iced. All the four samples were kept in a chilled room for storage studies. Sample C, chilled and stored in CSW, recorded a gradual gain in weight and an increase in salt content of the muscle. Presence of salt did not seem to cause any excessive protein denaturation. Salt extractability decreased at a gradual rate in all cases. Presence of salt seemed to wield no noticeable influence on lipid hydrolysis and subsequent peroxidation. Results of chemical and sensory evaluations highlight this. Holding sardines in CSW gave a product of excellent quality for the first four to five days of storage. Beyond the fifth day of storage quality deteriorated rapidly and there was no noticeable superiority for this sample (sample C) over the on board iced fish. This was evident in the sensory evaluation as well. However, a storage life of five days in a readily acceptable state is sufficient for the fish to be disposed in the market at a premium sale price over other landings of the same species.

Unsupervised intra-lingual and cross-lingual speaker adaptation for HMM-based speech synthesis using two-pass decision tree construction

Hidden Markov model (HMM)-based speech synthesis systems possess several advantages over concatenative synthesis systems. One such advantage is the relative ease with which HMM-based systems are adapted to speakers not present in the training dataset. Speaker adaptation methods used in the field of HMM-based automatic speech recognition (ASR) are adopted for this task. In the case of unsupervised speaker adaptation, previous work has used a supplementary set of acoustic models to estimate the transcription of the adaptation data. This paper first presents an approach to the unsupervised speaker adaptation task for HMM-based speech synthesis models which avoids the need for such supplementary acoustic models. This is achieved by defining a mapping between HMM-based synthesis models and ASR-style models, via a two-pass decision tree construction process. Second, it is shown that this mapping also enables unsupervised adaptation of HMM-based speech synthesis models without the need to perform linguistic analysis of the estimated transcription of the adaptation data. Third, this paper demonstrates how this technique lends itself to the task of unsupervised cross-lingual adaptation of HMM-based speech synthesis models, and explains the advantages of such an approach. Finally, listener evaluations reveal that the proposed unsupervised adaptation methods deliver performance approaching that of supervised adaptation.

Evaluating holonic control systems: A case study

in the last 10 years many designs and trial implementations of holonic manufacturing systems have been reported in the literature. Few of these have resulted in any industrial take up of the approach and part of this lack of adoption might be attributed to a shortage of evaluations of the resulting designs and implementations and their comparison with more conventional approaches. This paper proposes a simple approach for evaluating the effectiveness of a holonic system design, with particular focus on the ability of the system to support reconfiguration (in the face of change). A case study relating to a laboratory assembly system is provided to demonstrate the evaluation approach. Copyright © 2005 IFAC.

Parameter estimation for agenda-based user simulation

This paper presents an agenda-based user simulator which has been extended to be trainable on real data with the aim of more closely modelling the complex rational behaviour exhibited by real users. The train-able part is formed by a set of random decision points that may be encountered during the process of receiving a system act and responding with a user act. A sample-based method is presented for using real user data to estimate the parameters that control these decisions. Evaluation results are given both in terms of statistics of generated user behaviour and the quality of policies trained with different simulators. Compared to a handcrafted simulator, the trained system provides a much better fit to corpus data and evaluations suggest that this better fit should result in improved dialogue performance. © 2010 Association for Computational Linguistics.

Co-occurrence flow for pedestrian detection

The last few years have seen considerable progress in pedestrian detection. Recent work has established a combination of oriented gradients and optic flow as effective features although the detection rates are still unsatisfactory for practical use. This paper introduces a new type of motion feature, the co-occurrence flow (CoF). The advance is to capture relative movements of different parts of the entire body, unlike existing motion features which extract internal motion in a local fashion. Through evaluations on the TUD-Brussels pedestrian dataset, we show that our motion feature based on co-occurrence flow contributes to boost the performance of existing methods. © 2011 IEEE.

Robust filtering and smoothing with gaussian processes

We propose a principled algorithm for robust Bayesian filtering and smoothing in nonlinear stochastic dynamic systems when both the transition function and the measurement function are described by non-parametric Gaussian process (GP) models. GPs are gaining increasing importance in signal processing, machine learning, robotics, and control for representing unknown system functions by posterior probability distributions. This modern way of system identification is more robust than finding point estimates of a parametric function representation. Our principled filtering/smoothing approach for GP dynamic systems is based on analytic moment matching in the context of the forward-backward algorithm. Our numerical evaluations demonstrate the robustness of the proposed approach in situations where other state-of-the-art Gaussian filters and smoothers can fail. © 2011 IEEE.

Improving LVCSR system combination using neural network language model cross adaptation

State-of-the-art large vocabulary continuous speech recognition (LVCSR) systems often combine outputs from multiple subsystems developed at different sites. Cross system adaptation can be used as an alternative to direct hypothesis level combination schemes such as ROVER. The standard approach involves only cross adapting acoustic models. To fully exploit the complimentary features among sub-systems, language model (LM) cross adaptation techniques can be used. Previous research on multi-level n-gram LM cross adaptation is extended to further include the cross adaptation of neural network LMs in this paper. Using this improved LM cross adaptation framework, significant error rate gains of 4.0%-7.1% relative were obtained over acoustic model only cross adaptation when combining a range of Chinese LVCSR sub-systems used in the 2010 and 2011 DARPA GALE evaluations. Copyright © 2011 ISCA.

Inverse induction motor modelling of a submarine propulsion system

An inherent trade-off exists in simulation model development and employment: a trade-off between the level of detail simulated and the simulation models computational cost. It is often desirable to simulate a high level of detail to a high degree of accuracy. However, due to the nature of design optimisation, which requires a large number of design evaluations, the application of such simulation models can be prohibitively expensive. A induction motor modelling approache to reduce the computational cost while maintaining a high level of detail and accuracy in the final design is presented. © 2012 IEEE.

N-best error simulation for training spoken dialogue systems

A recent trend in spoken dialogue research is the use of reinforcement learning to train dialogue systems in a simulated environment. Past researchers have shown that the types of errors that are simulated can have a significant effect on simulated dialogue performance. Since modern systems typically receive an N-best list of possible user utterances, it is important to be able to simulate a full N-best list of hypotheses. This paper presents a new method for simulating such errors based on logistic regression, as well as a new method for simulating the structure of N-best lists of semantics and their probabilities, based on the Dirichlet distribution. Off-line evaluations show that the new Dirichlet model results in a much closer match to the receiver operating characteristics (ROC) of the live data. Experiments also show that the logistic model gives confusions that are closer to the type of confusions observed in live situations. The hope is that these new error models will be able to improve the resulting performance of trained dialogue systems. © 2012 IEEE.

Language model cross adaptation for LVCSR system combination

State-of-the-art large vocabulary continuous speech recognition (LVCSR) systems often combine outputs from multiple sub-systems that may even be developed at different sites. Cross system adaptation, in which model adaptation is performed using the outputs from another sub-system, can be used as an alternative to hypothesis level combination schemes such as ROVER. Normally cross adaptation is only performed on the acoustic models. However, there are many other levels in LVCSR systems' modelling hierarchy where complimentary features may be exploited, for example, the sub-word and the word level, to further improve cross adaptation based system combination. It is thus interesting to also cross adapt language models (LMs) to capture these additional useful features. In this paper cross adaptation is applied to three forms of language models, a multi-level LM that models both syllable and word sequences, a word level neural network LM, and the linear combination of the two. Significant error rate reductions of 4.0-7.1% relative were obtained over ROVER and acoustic model only cross adaptation when combining a range of Chinese LVCSR sub-systems used in the 2010 and 2011 DARPA GALE evaluations. © 2012 Elsevier Ltd. All rights reserved.

Active learning of model evidence using Bayesian quadrature

Numerical integration is a key component of many problems in scientific computing, statistical modelling, and machine learning. Bayesian Quadrature is a modelbased method for numerical integration which, relative to standard Monte Carlo methods, offers increased sample efficiency and a more robust estimate of the uncertainty in the estimated integral. We propose a novel Bayesian Quadrature approach for numerical integration when the integrand is non-negative, such as the case of computing the marginal likelihood, predictive distribution, or normalising constant of a probabilistic model. Our approach approximately marginalises the quadrature model's hyperparameters in closed form, and introduces an active learning scheme to optimally select function evaluations, as opposed to using Monte Carlo samples. We demonstrate our method on both a number of synthetic benchmarks and a real scientific problem from astronomy.

Phylogenetic relationships of Danio within the order cypriniformes: A framework for comparative and evolutionary studies of a model species

The evolutionary relationships of species of Danio and the monophyly and phylogenetic placement of the genus within the family Cyprinidae and subfamily Rasborinae provide fundamentally important phyloinformatics necessary for direct evaluations of an array of pertinent questions in modern comparative biology. Although the genus Danio is not one of the most diverse within the family, Danio rerio is one of the most important model species in biology. Many investigations have used this species or presumed close relatives to address specific questions that have lasting impact on the hypothesis and theory of development in vertebrates. Largely lacking from this approach has been a holistic picture of the exact phylogenetic or evolutionary relationships of this species and its close relatives. One thing that has been learned over the previous century is that many organismal attributes (e.g., developmental pathways, ecologies, behaviors, speciation) are historically constrained and their origins and functions are best explained via a phylogenetic approach. Herein, we provide a molecular evaluation of the phylogenetic placement of the model species Danio rerio within the genus Danio and among hypothesized closely related species and genera. Our analysis is derived from data using two nuclear genes (RAG1, rhodopsin) and five mitochondrial genes (ND4, ND4L, ND5, COI, cyt b) evaluated using parsimony, maximum likelihood, and Bayesian analyses. The family Cyprinidae is resolved as monophyletic but the subfamily Rasborinae (priority over Danioinae) is an unnatural assemblage. Danio is identified as a monophyletic group sister to a clade inclusive of the genera Chela, Microrasbora, Devario, and Inlecypris, not Devario nor Esomus as hypothesized in previous studies. Danio rerio is sister to D. kyathit among the species of Danio evaluated in this analysis. Microrasbora and Rasbora are non-monophyletic assemblages; however, Boraras is monophyletic.