Investigating fuzzy-input fuzzy-output support vector machines for robust voice quality classification

The dynamic use of voice qualities in spoken language can reveal useful information on a speakers attitude, mood and affective states. This information may be very desirable for a range of, both input and output, speech technology applications. However, voice quality annotation of speech signals may frequently produce far from consistent labeling. Groups of annotators may disagree on the perceived voice quality, but whom should one trust or is the truth somewhere in between? The current study looks first to describe a voice quality feature set that is suitable for differentiating voice qualities on a tense to breathy dimension. Further, the study looks to include these features as inputs to a fuzzy-input fuzzy-output support vector machine (F²SVM) algorithm, which is in turn capable of softly categorizing voice quality recordings. The F²SVM is compared in a thorough analysis to standard crisp approaches and shows promising results, while outperforming for example standard support vector machines with the sole difference being that the F²SVM approach receives fuzzy label information during training. Overall, it is possible to achieve accuracies of around 90% for both speaker dependent (cross validation) and speaker independent (leave one speaker out validation) experiments. Additionally, the approach using F²SVM performs at an accuracy of 82% for a cross corpus experiment (i.e. training and testing on entirely different recording conditions) in a frame-wise analysis and of around 97% after temporally integrating over full sentences. Furthermore, the output of fuzzy measures gave performances close to that of human annotators.     

Working-memory capacity, proactive interference, and divided attention: Limits on long-term memory retrieval.

Two experiments examined how individual differences in working-memory capacity (WM) relate to proactive interference (PI) susceptibility. We tested high and low WM-span participants in a PI-buildup task under single-task or dual-task ("load") conditions. In Experiment 1, a finger-tapping task was imposed during encoding and retrieval of each list; in Experiment 2, tapping was required during encoding or retrieval. In both experiments, low spans showed greater PI than did high spans under no load, but groups showed equivalent PI under divided attention. Load increased PI only for high spans, suggesting they use attention at encoding and retrieval to combat PI. In Experiment 2, only low spans showed a dual-task cost on List 1 memory, before PI built up. Results indicate a role for attentional processing, perhaps inhibitory in nature, at encoding and retrieval, and are discussed with respect to theories of WM and prefrontal cortex function. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The apolipoprotein B R3531C mutation. Characteristics of 24 subjects from 9 kindreds

Familial ligand-defective apolipoprotein B (apoB) is a group of disorders caused by mutations in the apoB gene. In this report the R3531C mutation is characterized further using a monoclonal antibody MB19/dynamic laser light scattering technique to measure ratios of Cys(3531) to normal low density lipoprotein (LDL) particles. All six subjects studied showed a preferential accumulation of particles carrying the defective apoB allotype. We determined binding properties of LDL from R3531C heterozygotes by measurement of high-affinity binding to LDL receptors on fibroblasts and its ability promote growth of U937 cells. LDL from R3531C heterozygotes, compared to normal LDL, had 49.3% of the binding affinity and was 74% as effective in a U937 cell proliferation assay. To identify new probands, we screened 2570 subjects for the R3531C mutation. Nine probands were found with 15 affected relatives. Of the seven haplotypes we uncovered, two were novel, while five were identical to one initially reported as associated with Cys3531. Three silent mutations were detected also: T3540T, N3542N and T3552T. Analysis of lipid profiles of R3531C families showed, as with the R3500Q mutation, variable expression of the phenotype, modulated by environmental and other genetic factors. Both mutations tend to produce lower plasma levels of LDL in affected subjects than do defects of the LDL receptor (familial hypercholesterolemia, FH). This study shows that the Cys(3531) LDL particles are not only defective at binding to the LDL receptor, as determined by two separate methods, but that in all cases they accumulate preferentially compared to the normal allotype.-Pullinger, C. R., D. Gaffney, M. M. Gutierrez, M. J. Malloy, V. N. Schumaker, C. J. Packard, and J. P. Kane. Apolipoprotein B R3531C mutation: characteristics of 24 subjects from 9 kindreds. .