Hidden Conditional Random Fields for Visual Speech Recognition

In this paper we present the application of Hidden Conditional Random Fields (HCRFs) to modelling speech for visual speech recognition. HCRFs may be easily adapted to model long range dependencies across an observation sequence. As a result visual word recognition performance can be improved as the model is able to take more of a contextual approach to generating state sequences. Results are presented from a speaker-dependent, isolated digit, visual speech recognition task using comparisons with a baseline HMM system. We firstly illustrate that word recognition rates on clean video using HCRFs can be improved by increasing the number of past and future observations being taken into account by each state. Secondly we compare model performances using various levels of video compression on the test set. As far as we are aware this is the first attempted use of HCRFs for visual speech recognition.

Passing quantum correlations to qubits using any two-mode state

We draw an explicit connection between the statistical properties of an entangled two-mode continuous variable (CV) resource and the amount of entanglement that can be dynamically transferred to a pair of noninteracting two-level systems. More specifically, we rigorously reformulate entanglement-transfer process by making use of covariance matrix formalism. When the resource state is Gaussian, our method makes the approach to the transfer of quantum correlations much more flexible than in previously considered schemes and allows the straightforward inclusion of the effects of noise affecting the CV system. Moreover, the proposed method reveals that the use of de-Gaussified two-mode states is almost never advantageous for transferring entanglement with respect to the full Gaussian picture, despite the entanglement in the non-Gaussian resource can be much larger than in its Gaussian counterpart. We can thus conclude that the entanglement-transfer map overthrows the

Phenotype-environment correlations in a putative whitefish adaptive radiation

1. The adaptive radiation of fishes into benthic (littoral) and pelagic (lentic) morphs in post-glaciallakes has become an important model system for speciation. Although these systems are well stud-ied, there is little evidence of the existence of morphs that have diverged to utilize resources in theremaining principal lake habitat, the profundal zone.
2. Here, we tested phenotype-environment correlations of three whitefish (Coregonus lavaretus)morphs that have radiated into littoral, pelagic and profundal niches in northern Scandinavianlakes. We hypothesized that morphs in such trimorphic systems would have a morphology adaptedto one of the principal lake habitats (littoral, pelagic or profundal niches). Most whitefish popula-tions in the study area are formed by a single (monomorphic) whitefish morph, and we furtherhypothesized that these populations should display intermediate morphotypes and niche utiliza-tion. We used a combination of traditional (stomach content, habitat use, gill raker counts) andmore recently developed (stable isotopes, geometric morphometrics) techniques to evaluate pheno-type-environment correlations in two lakes with trimorphic and two lakes with monomorphicwhitefish.
3. Distinct phenotype-environment correlations were evident for each principal niche in whitefishmorphs inhabiting trimorphic lakes. Monomorphic whitefish exploited multiple habitats, hadintermediate morphology, displayed increased variance in gillraker-counts, and relied significantlyon zooplankton, most likely due to relaxed resource competition.
4. We suggest that the ecological processes acting in the trimorphic lakes are similar to each other,and are driving the adaptive evolution of whitefish morphs, possibly leading to the formation ofnew species.

Assessment of the benefits of Discrete Conditional Survival Models in modelling ambulance response times

Many of the challenges faced in health care delivery can be informed through building models. In particular, Discrete Conditional Survival (DCS) models, recently under development, can provide policymakers with a flexible tool to assess time-to-event data. The DCS model is capable of modelling the survival curve based on various underlying distribution types and is capable of clustering or grouping observations (based on other covariate information) external to the distribution fits. The flexibility of the model comes through the choice of data mining techniques that are available in ascertaining the different subsets and also in the choice of distribution types available in modelling these informed subsets. This paper presents an illustrated example of the Discrete Conditional Survival model being deployed to represent ambulance response-times by a fully parameterised model. This model is contrasted against use of a parametric accelerated failure-time model, illustrating the strength and usefulness of Discrete Conditional Survival models.

Costs of resistance: genetic correlations and potential trade-offs in an insect immune system

Theory predicts that natural selection will erode additive genetic variation in fitness-related traits. However, numerous studies have found considerable heritable variation in traits related to immune function, which should be closely linked to fitness. This could be due to trade-offs maintaining variation in these traits. We used the Egyptian cotton leafworm, Spodoptera littoralis, as a model system to examine the quantitative genetics of insect immune function. We estimated the heritabilities of several different measures of innate immunity and the genetic correlations between these immune traits and a number of life history traits. Our results provide the first evidence for a potential genetic trade-off within the insect immune system, with antibacterial activity (lysozyme-like) exhibiting a significant negative genetic correlation with haemocyte density, which itself is positively genetically correlated with both haemolymph phenoloxidase activity and cuticular melanization. We speculate on a potential trade-off between defence against parasites and predators, mediated by larval colour, and its role in maintaining genetic variation in traits under natural selection.

Discrete Conditional Phase-type models utilising classification trees: Application to modelling health service capacities

Discrete Conditional Phase-type (DC-Ph) models consist of a process component (survival distribution) preceded by a set of related conditional discrete variables. This paper introduces a DC-Ph model where the conditional component is a classification tree. The approach is utilised for modelling health service capacities by better predicting service times, as captured by Coxian Phase-type distributions, interfaced with results from a classification tree algorithm. To illustrate the approach, a case-study within the healthcare delivery domain is given, namely that of maternity services. The classification analysis is shown to give good predictors for complications during childbirth. Based on the classification tree predictions, the duration of childbirth on the labour ward is then modelled as either a two or three-phase Coxian distribution. The resulting DC-Ph model is used to calculate the number of patients and associated bed occupancies, patient turnover, and to model the consequences of changes to risk status.

Distributing fully optomechanical quantum correlations

We present a scheme to prepare quantum correlated states of two mechanical systems based on the pouring of preavailable all-optical entanglement into the state of two micromirrors belonging to remote and noninteracting optomechanical cavities. We show that, under realistic experimental conditions, the protocol allows for the preparation of a genuine quantum state of a composite mesoscopic system whose nonclassical features extend beyond the occurrence of entanglement. We finally discuss a way to access such mechanical correlations.

Theoretical and experimental correlations of gas dissolution, diffusion, and thermodynamic properties in determination of gas permeability and selectivity in supported ionic liquid membranes

Supported ionic liquid membranes (SILMs) has the potential to be a new technological platform for gas/organic vapour separation because of the unique non-volatile nature and discriminating gas dissolution properties of room temperature ionic liquids (ILs). This work starts with an examination of gas dissolution and transport properties in bulk imidazulium cation based ionic liquids [C_nmim][NTf2] (n = 2.4, 6, 8.10) from simple gas H₂, N₂, to polar CO₂, and C₂H₆, leading to a further analysis of how gas dissolution and diffusion are influenced by molecular specific gas-SILMs interactions, reflected by differences in gas dissolution enthalpy and entropy. These effects were elucidated again during gas permeation studies by examining how changes in these properties and molecular specific interactions work together to cause deviations from conventional solution–diffusion theory and their impact on some remarkably contrasting gas perm-selectivity performance. The experimental perm-selectivity for all tested gases showed varied and contrasting deviation from the solution–diffusion, depending on specific gas-IL combinations. It transpires permeation for simpler non-polar gases (H₂, N₂) is diffusion controlled, but strong molecular specific gas-ILs interactions led to a different permeation and selectivity performance for C₂H₆ and CO₂. With exothermic dissolution enthalpy and large order disruptive entropy, C₂H₆ displayed the fastest permeation rate at increased gas phase pressure in spite of its smallest diffusivity among the tested gases. The C₂H₆ gas molecules “peg” on the side alkyl chain on the imidazulium cation at low concentration, and are well dispersed in the ionic liquids phase at high concentration. On the other hand strong CO₂-ILs affinity resulted in a more prolonged “residence time” for the gas molecule, typified by reversed CO₂/N₂ selectivity and slowest CO₂ transport despite CO₂ possess the highest solubility and comparable diffusivity in the ionic liquids. The unique transport and dissolution behaviour of CO₂ are further exploited by examining the residing state of CO₂ molecules in the ionic liquid phase, which leads to a hypothesis of a condensing and holding capacity of ILs towards CO₂, which provide an explanation to slower CO₂ transport through the SILMs. The pressure related exponential increase in permeations rate is also analysed which suggests a typical concentration dependent diffusion rate at high gas concentration under increased gas feed pressure. Finally the strong influence of discriminating and molecular specific gas-ILs interactions on gas perm-selectivity performance points to future specific design of ionic liquids for targeted gas separations.

Faithful nonclassicality indicators and extremal quantum correlations in two-qubit states

The state disturbance induced by locally measuring a quantum system yields a signature of nonclassical correlations beyond entanglement. Here, we present a detailed study of such correlations for two-qubit mixed states. To overcome the asymmetry of quantum discord and the unfaithfulness of measurement-induced disturbance (severely overestimating quantum correlations), we propose an ameliorated measurement-induced disturbance as nonclassicality indicator, optimized over joint local measurements, and we derive its closed expression for relevant two-qubit states. We study its analytical relation with discord, and characterize the maximally quantum-correlated mixed states, that simultaneously extremize both quantifiers at given von Neumann entropy: among all two-qubit states, these states possess the most robust quantum correlations against noise.

Extremal quantum correlations: Experimental study with two-qubit states

We explore experimentally the space of two-qubit quantum-correlated mixed states, including frontier states as defined by the use of quantum discord and von Neumann entropy. Our experimental setup is flexible enough to allow for high-quality generation of a vast variety of states. We address quantitatively the relation between quantum discord and a recently suggested alternative measure of quantum correlations.