Bayesian adaptation and adaptively trained systems

As the use of found data increases, more systems are being built using adaptive training. Here transforms are used to represent unwanted acoustic variability, e.g. speaker and acoustic environment changes, allowing a canonical model that models only the "pure" variability of speech to be trained. Adaptive training may be described within a Bayesian framework. By using complexity control approaches to ensure robust parameter estimates, the standard point estimate adaptive training can be justified within this Bayesian framework. However during recognition there is usually no control over the amount of data available. It is therefore preferable to be able to use a full Bayesian approach to applying transforms during recognition rather than the standard point estimates. This paper discusses various approximations to Bayesian approaches including a new variational Bayes approximation. The application of these approaches to state-of-the-art adaptively trained systems using both CAT and MLLR transforms is then described and evaluated on a large vocabulary speech recognition task. © 2005 IEEE.

Dog chromosome-specific paints reveal evolutionary inter- and intrachromosomal rearrangements in the American mink and human

Forty chromosome-specific paint probes of the domestic dog (Canis familiaris, 2n = 78) were used to delineate conserved segments on metaphase chromosomes of the American mink (Mustela vison, 2n = 30) by fluorescence in situ hybridisation. Half of the 38 canine autosomal probes each painted one pair of homologous segments in a diploid mink metaphase, whereas the other 19 dog probes each painted from two to five pairs of discrete segments. In total, 38 canine autosomal paints highlighted 71 pairs of conserved segments in the mink. These painting results allow us to establish a complete comparative chromosome map between the American mink and domestic dog. This map demonstrates that extensive chromosome rearrangements differentiate the karyotypes of the dog and American mink. The 38 dog autosomes could be reconstructed from the 14 autosomes of the American mink through at least 47 fissions, 25 chromosome fusions, and six inversions. Furthermore, comparison of the current dog/mink map with the published human/dog map discloses 23 cryptic intrachromosomal rearrangements in 10 regions of conserved synteny in the human and American mink genomes and thus further refined the human/mink comparative genome map. Copyright (C) 2000 S. Karger AG, Basel.

Comparison of Pax1/9 locus reveals 500-Myr-old syntenic block and evolutionary conserved noncoding regions

Identification of conserved genomic regions within and between different genomes is crucial when studying genome evolution. Here, we described regions of strong synteny conservation between vertebrate deuterostomes (tetrapods and teleosts) and invertebrat

Constrained discriminative mapping transforms for unsupervised speaker adaptation

Discriminative mapping transforms (DMTs) is an approach to robustly adding discriminative training to unsupervised linear adaptation transforms. In unsupervised adaptation DMTs are more robust to unreliable transcriptions than directly estimating adaptation transforms in a discriminative fashion. They were previously proposed for use with MLLR transforms with the associated need to explicitly transform the model parameters. In this work the DMT is extended to CMLLR transforms. As these operate in the feature space, it is only necessary to apply a different linear transform at the front-end rather than modifying the model parameters. This is useful for rapidly changing speakers/environments. The performance of DMTs with CMLLR was evaluated on the WSJ 20k task. Experimental results show that DMTs based on constrained linear transforms yield 3% to 6% relative gain over MLE transforms in unsupervised speaker adaptation. © 2011 IEEE.

Multiple object tracking using evolutionary MCMC-based particle algorithms

Algorithms are presented for detection and tracking of multiple clusters of co-ordinated targets. Based on a Markov chain Monte Carlo sampling mechanization, the new algorithms maintain a discrete approximation of the filtering density of the clusters' state. The filters' tracking efficiency is enhanced by incorporating various sampling improvement strategies into the basic Metropolis-Hastings scheme. Thus, an evolutionary stage consisting of two primary steps is introduced: 1) producing a population of different chain realizations, and 2) exchanging genetic material between samples in this population. The performance of the resulting evolutionary filtering algorithms is demonstrated in two different settings. In the first, both group and target properties are estimated whereas in the second, which consists of a very large number of targets, only the clustering structure is maintained. © 2009 IFAC.

Seasonal changes in the trade-off among fig-supported wasps and viable seeds in figs and their evolutionary implications

What the real trade-off is among fig-supported wasps and the viable seeds of figs is heatedly debated in the studies of fig/fig wasp mutualism. In the present study, we collected wasp offspring (galls) and the viable seeds of premature fruits, and determined the foundress number in receptive fruits and all the types of wasps supported by Ficus racemosa L. during both the rainy and dry seasons in Xishuangbanna, China. The data show that the galls were positively correlated with viable seeds (n=32;r=0.74; P < 0.001) when the proportion of vacant female flowers (PVFF) was high, in April (68.0%), and were negatively correlated with viable seeds (n=48;r=-0.59; P < 0.05) when PVFF were limited (PVFF 42.6%) during a colder month (January). The mean foundress number per fruit during the colder months is significantly lower than during the warmer months (F-5,F-603 = 27.9; P < 0.001) and pollinator wasps can live longer during the colder months, During the colder months, the proportions of non-pollinators and wasp offspring are higher than those found during other months, whereas the proportion of viable seeds is not different compared with that of other months. Non-pollinator wasps tend to oviposit the female flowers that have been oviposited by pollinator wasps. The non-pollinators only negatively affect pollinator wasps and there is no obvious negative effect of non-pollinator wasps on viable seeds, so ovipositing by non-pollinator wasps will not result in the extinction of the figs during the process of evolution. The results of the present study indicate that figs can allow less foundresses to be in fruit cavities when PVFF are limited, which provides supporting evidence for the previous assumption that the plants have developed a mechanism to maintain a stable system because of the conflicts between the parties involved.

A single-rate context-dependent learning process underlies rapid adaptation to familiar object dynamics.

Motor learning has been extensively studied using dynamic (force-field) perturbations. These induce movement errors that result in adaptive changes to the motor commands. Several state-space models have been developed to explain how trial-by-trial errors drive the progressive adaptation observed in such studies. These models have been applied to adaptation involving novel dynamics, which typically occurs over tens to hundreds of trials, and which appears to be mediated by a dual-rate adaptation process. In contrast, when manipulating objects with familiar dynamics, subjects adapt rapidly within a few trials. Here, we apply state-space models to familiar dynamics, asking whether adaptation is mediated by a single-rate or dual-rate process. Previously, we reported a task in which subjects rotate an object with known dynamics. By presenting the object at different visual orientations, adaptation was shown to be context-specific, with limited generalization to novel orientations. Here we show that a multiple-context state-space model, with a generalization function tuned to visual object orientation, can reproduce the time-course of adaptation and de-adaptation as well as the observed context-dependent behavior. In contrast to the dual-rate process associated with novel dynamics, we show that a single-rate process mediates adaptation to familiar object dynamics. The model predicts that during exposure to the object across multiple orientations, there will be a degree of independence for adaptation and de-adaptation within each context, and that the states associated with all contexts will slowly de-adapt during exposure in one particular context. We confirm these predictions in two new experiments. Results of the current study thus highlight similarities and differences in the processes engaged during exposure to novel versus familiar dynamics. In both cases, adaptation is mediated by multiple context-specific representations. In the case of familiar object dynamics, however, the representations can be engaged based on visual context, and are updated by a single-rate process.