Algorithms for Hidden Markov Models With Imprecisely Specified Parameters

Hidden Markov models (HMMs) are widely used models for sequential data. As with other probabilistic graphical models, they require the specification of precise probability values, which can be too restrictive for some domains, especially when data are scarce or costly to acquire. We present a generalized version of HMMs, whose quantification can be done by sets of, instead of single, probability distributions. Our models have the ability to suspend judgment when there is not enough statistical evidence, and can serve as a sensitivity analysis tool for standard non-stationary HMMs. Efficient inference algorithms are developed to address standard HMM usage such as the computation of likelihoods and most probable explanations. Experiments with real data show that the use of imprecise probabilities leads to more reliable inferences without compromising efficiency.

Hidden Markov Models With Set-Valued Parameters

Hidden Markov models (HMMs) are widely used probabilistic models of sequential data. As with other probabilistic models, they require the specification of local conditional probability distributions, whose assessment can be too difficult and error-prone, especially when data are scarce or costly to acquire. The imprecise HMM (iHMM) generalizes HMMs by allowing the quantification to be done by sets of, instead of single, probability distributions. iHMMs have the ability to suspend judgment when there is not enough statistical evidence, and can serve as a sensitivity analysis tool for standard non-stationary HMMs. In this paper, we consider iHMMs under the strong independence interpretation, for which we develop efficient inference algorithms to address standard HMM usage such as the computation of likelihoods and most probable explanations, as well as performing filtering and predictive inference. Experiments with real data show that iHMMs produce more reliable inferences without compromising the computational efficiency.

Inter-Frame Contextual Modelling For Visual Speech Recognition

In this paper, we present a new approach to visual speech recognition which improves contextual modelling by combining Inter-Frame Dependent and Hidden Markov Models. This approach captures contextual information in visual speech that may be lost using a Hidden Markov Model alone. We apply contextual modelling to a large speaker independent isolated digit recognition task, and compare our approach to two commonly adopted feature based techniques for incorporating speech dynamics. Results are presented from baseline feature based systems and the combined modelling technique. We illustrate that both of these techniques achieve similar levels of performance when used independently. However significant improvements in performance can be achieved through a combination of the two. In particular we report an improvement in excess of 17% relative Word Error Rate in comparison to our best baseline system.

FPGA Implementation of a Pipelined Gaussian Calculation for HMM-Based Large Vocabulary Speech Recognition

A scalable large vocabulary, speaker independent speech recognition system is being developed using Hidden Markov Models (HMMs) for acoustic modeling and a Weighted Finite State Transducer (WFST) to compile sentence, word, and phoneme models. The system comprises a software backend search and an FPGA-based Gaussian calculation which are covered here. In this paper, we present an efficient pipelined design implemented both as an embedded peripheral and as a scalable, parallel hardware accelerator. Both architectures have been implemented on an Alpha Data XRC-5T1, reconfigurable computer housing a Virtex 5 SX95T FPGA. The core has been tested and is capable of calculating a full set of Gaussian results from 3825 acoustic models in 9.03 ms which coupled with a backend search of 5000 words has provided an accuracy of over 80%. Parallel implementations have been designed with up to 32 cores and have been successfully implemented with a clock frequency of 133?MHz.

Combining missing-feature theory, speech enhancement, and speaker-dependent/-independent modeling for speech separation

This paper considers the separation and recognition of overlapped speech sentences assuming single-channel observation. A system based on a combination of several different techniques is proposed. The system uses a missing-feature approach for improving crosstalk/noise robustness, a Wiener filter for speech enhancement, hidden Markov models for speech reconstruction, and speaker-dependent/-independent modeling for speaker and speech recognition. We develop the system on the Speech Separation Challenge database, involving a task of separating and recognizing two mixing sentences without assuming advanced knowledge about the identity of the speakers nor about the signal-to-noise ratio. The paper is an extended version of a previous conference paper submitted for the challenge.

Probabilistic Inference in Credal Networks: New Complexity Results

Credal networks are graph-based statistical models whose parameters take values in a set, instead of being sharply specified as in traditional statistical models (e.g., Bayesian networks). The computational complexity of inferences on such models depends on the irrelevance/independence concept adopted. In this paper, we study inferential complexity under the concepts of epistemic irrelevance and strong independence. We show that inferences under strong independence are NP-hard even in trees with binary variables except for a single ternary one. We prove that under epistemic irrelevance the polynomial-time complexity of inferences in credal trees is not likely to extend to more general models (e.g., singly connected topologies). These results clearly distinguish networks that admit efficient inferences and those where inferences are most likely hard, and settle several open questions regarding their computational complexity. We show that these results remain valid even if we disallow the use of zero probabilities. We also show that the computation of bounds on the probability of the future state in a hidden Markov model is the same whether we assume epistemic irrelevance or strong independence, and we prove an analogous result for inference in Naive Bayes structures. These inferential equivalences are important for practitioners, as hidden Markov models and Naive Bayes networks are used in real applications of imprecise probability.

Combining global climate and regional landscape models to improve prediction of invasion risk

Aim
It is widely acknowledged that species distributions result from a variety of biotic and abiotic factors operating at different spatial scales. Here, we aimed to (1) determine the extent to which global climate niche models (CNMs) can be improved by the addition of fine-scale regional data; (2) examine climatic and environmental factors influencing the range of 15 invasive aquatic plant species; and (3) provide a case study for the use of such models in invasion management on an island.

Location
Global, with a case study of species invasions in Ireland.

Methods
Climate niche models of global extent (including climate only) and regional environmental niche models (with additional factors such as human influence, land use and soil characteristics) were generated using maxent for 15 invasive aquatic plants. The performance of these models within the invaded range of the study species in Ireland was assessed, and potential hotspots of invasion suitability were determined. Models were projected forward up to 2080 based on two climate scenarios.

Results
While climate variables are important in defining the global range of species, factors related to land use and nutrient level were of greater importance in regional projections. Global climatic models were significantly improved at the island scale by the addition of fine-scale environmental variables (area under the curve values increased by 0.18 and true skill statistic values by 0.36), and projected ranges decreased from an average of 86% to 36% of the island.

Main conclusions
Refining CNMs with regional data on land use, human influence and landscape may have a substantial impact on predictive capacity, providing greater value for prioritization of conservation management at subregional or local scales.

Effect of major lifestyle risk factors, independent and jointly, on life expectancy with and without cardiovascular disease: results from the Consortium on Health and Ageing Network of Cohorts in Europe and the United States (CHANCES)

Seldom have studies taken account of changes in lifestyle habits in the elderly, or investigated their impact on disease-free life expectancy (LE) and LE with cardiovascular disease (CVD). Using data on subjects aged 50+ years from three European cohorts (RCPH, ESTHER and Tromsø), we used multi-state Markov models to calculate the independent and joint effects of smoking, physical activity, obesity and alcohol consumption on LE with and without CVD. Men and women aged 50 years who have a favourable lifestyle (overweight but not obese, light/moderate drinker, non-smoker and participates in vigorous physical activity) lived between 7.4 (in Tromsø men) and 15.7 (in ESTHER women) years longer than those with an unfavourable lifestyle (overweight but not obese, light/moderate drinker, smoker and does not participate in physical activity). The greater part of the extra life years was in terms of "disease-free" years, though a healthy lifestyle was also associated with extra years lived after a CVD event. There are sizeable benefits to LE without CVD and also for survival after CVD onset when people favour a lifestyle characterized by salutary behaviours. Remaining a non-smoker yielded the greatest extra years in overall LE, when compared to the effects of routinely taking physical activity, being overweight but not obese, and drinking in moderation. The majority of the overall LE benefit is in disease free years. Therefore, it is important for policy makers and the public to know that prevention through maintaining a favourable lifestyle is "never too late".

Testing genuine multipartite nonlocality in phase spacea

We demonstrate genuine three-mode nonlocality based on phase-space formalism. A Svetlichny-type Bell inequality is formulated in terms of the s-parametrized quasiprobability function. We test such a tool using exemplary forms of three-mode entangled states, identifying the ideal measurement settings required for each state. We thus verify the presence of genuine three-mode nonlocality that cannot be reproduced by local or nonlocal hidden variable models between any two out of three modes. In our results, GHZ- and W-type nonlocality can be fully discriminated. We also study the behavior of genuine tripartite nonlocality under the effects of detection inefficiency and dissipation induced by local thermal environments. Our formalism can be useful to test the sharing of genuine multipartite quantum correlations among the elements of some interesting physical settings, including arrays of trapped ions and intracavity ultracold atoms. DOI: 10.1103/PhysRevA.87.022123

Synthetic light curves and spectra for three-dimensional delayed-detonation models of type ia supernovae

In a companion paper, Seitenzahl et al. have presented a set of three-dimensional delayed detonation models for thermonuclear explosions of near-Chandrasekhar-mass white dwarfs (WDs). Here,we present multidimensional radiative transfer simulations that provide synthetic light curves and spectra for those models. The model sequence explores both changes in the strength of the deflagration phase (which is controlled by the ignition configuration in our models) and the WD central density. In agreement with previous studies, we find that the strength of the deflagration significantly affects the explosion and the observables. Variations in the central density also have an influence on both brightness and colour, but overall it is a secondary parameter in our set of models. In many respects, the models yield a good match to the observed properties of normal Type Ia supernovae (SNe Ia): peak brightness, rise/decline time-scales and synthetic spectra are all in reasonable agreement. There are, however, several differences. In particular, the models are systematically too red around maximum light, manifest spectral line velocities that are a little too high and yield I-band light curves that do not match observations. Although some of these discrepancies may simply relate to approximations made in the modelling, some pose real challenges to the models. If viewed as a complete sequence, our models do not reproduce the observed light-curve width- luminosity relation (WLR) of SNe Ia: all our models show rather similar B-band decline rates, irrespective of peak brightness. This suggests that simple variations in the strength of the deflagration phase in Chandrasekhar-mass deflagration-to-detonation models do not readily explain the observed diversity of normal SNe Ia. This may imply that some other parameter within the Chandrasekhar-mass paradigm is key to the WLR, or that a substantial fraction of normal SNe Ia arise from an alternative explosion scenario.

Synthesizable high performance adaptive equaliser and viterbi decoder for the class-IV PRML channel

The design and VLSI implementation of two key components of the class-IV partial response maximum likelihood channel (PR-IV) the adaptive filter and the Viterbi decoder are described. These blocks are implemented using parameterised VHDL modules, from a library of common digital signal processing (DSP) and arithmetic functions. Design studies, based on 0.6 micron 3.3V standard cell processes, indicate that worst case sampling rates of 49 mega-samples per second are achievable for this system, with proportionally high sampling rates for full custom designs and smaller dimension processes. Significant increases in the sampling rate, from 49 MHz to approximately 180 MHz, can be achieved by operating four filter modules in parallel, and this implementation has 50% lower power consumption than a pipelined filter operating at the same speed.

Kernel based approaches to local nonlinear non-parametric variable selection

In this paper, we consider the variable selection problem for a nonlinear non-parametric system. Two approaches are proposed, one top-down approach and one bottom-up approach. The top-down algorithm selects a variable by detecting if the corresponding partial derivative is zero or not at the point of interest. The algorithm is shown to have not only the parameter but also the set convergence. This is critical because the variable selection problem is binary, a variable is either selected or not selected. The bottom-up approach is based on the forward/backward stepwise selection which is designed to work if the data length is limited. Both approaches determine the most important variables locally and allow the unknown non-parametric nonlinear system to have different local dimensions at different points of interest. Further, two potential applications along with numerical simulations are provided to illustrate the usefulness of the proposed algorithms.