Graph-based Semi-Supervised Learning in Acoustic Modeling for Automatic Speech Recognition

Thesis (Ph.D.)--University of Washington, 2016-06

Modelling the structure of latexin-carboxypeptidase A complex based on chemical cross-linking and molecular docking

We have determined the three-dimensional structure of the protein complex between latexin and carboxypeptidase A using a combination of chemical cross-linking, mass spectrometry and molecular docking. The locations of three intermolecular cross-links were identified using mass spectrometry and these constraints were used in combination with a speed-optimised docking algorithm allowing us to evaluate more than 3 x 10(11) possible conformations. While cross-links represent only limited structural constraints, the combination of only three experimental cross-links with very basic molecular docking was sufficient to determine the complex structure. The crystal structure of the complex between latexin and carboxypeptidase A4 determined recently allowed us to assess the success of this structure determination approach. Our structure was shown to be within 4 angstrom r.m.s. deviation of C alpha atoms of the crystal structure. The study demonstrates that cross-linking in combination with mass spectrometry can lead to efficient and accurate structural modelling of protein complexes.

Recognition and inactivation of LPS by lipophorin particles

Lipophorin is the major lipid carrier in insects, but various observations indicate that lipophorin is also involved in immune reactions. To examine a possible role of lipophorin in defence reactions, we mixed hemolymph plasma from Galleria mellonella with LPS and noticed that lipophorin forms detergent-insoluble aggregates, while most other plasma proteins are not affected. Lipophorin particles isolated by low-density gradient centrifugation retained LPS-induced aggregation properties, which suggested to us that these immune-reactive particles are able to recognise LPS and respond by forming insoluble aggregates. Antibodies against LPS-binding proteins, such as immulectin-2 and beta-1,3-glucan binding protein, cross-reacted with proteins associated with purified lipophorin particles. To examine whether LPS-mediated aggregates inactivate LPS, we added LPS-lipophorin mixtures to purified lipophorin particles and monitored aggregate formation. Under these conditions lipophorin did not form insoluble aggregates, which indicates that lipophorin particles sequester LPS into non-toxic aggregates. (c) 2005 Elsevier Ltd. All rights reserved.

Novel monoclonal antibody recognition of oxidative DNA damage adduct, deoxycytidine-glyoxal

Glyoxal, a reactive aldehyde, is a decomposition product of lipid hydroperoxides, oxidative deoxyribose breakdown, or autoxidation of sugars, such as glucose. It readily forms DNA adducts, generating potential carcinogens such as glyoxalated deoxycytidine (gdC). A major drawback in assessing gdC formation in cellular DNA has been methodologic sensitivity. We have developed an mAb that specifically recognizes gdC. Balb/c mice were immunized with DNA, oxidatively modified by UVC/hydrogen peroxide in the presence of endogenous metal ions. Although UVC is not normally considered an oxidizing agent, a UVC/hydrogen peroxide combination may lead to glyoxalated bases arising from hydroxyl radical damage to deoxyribose. This damaging system was used to induce numerous oxidative lesions including glyoxal DNA modifications, from which resulted a number of clones. Clone F3/9/H2/G5 showed increased reactivity toward glyoxal-modified DNA greater than that of the immunizing antigen. ELISA unequivocally showed Ab recognition toward gdC, which was confirmed by gas chromatography-mass spectrometry of the derivatized adduct after formic acid hydrolysis to the modified base. Binding of Ab F3/9 with glyoxalated and untreated oligomers containing deoxycytidine, deoxyguanosine, thymidine, and deoxyadenosine assessed by ELISA produced significant recognition (p 0.0001) of glyoxal-modified deoxycytidine greater than that of untreated oligomer. Additionally, inhibition ELISA studies using the glyoxalated and native deoxycytidine oligomer showed increased recognition for gdC with more than a 5-fold difference in IC50 values. DNA modified with increasing levels of iron (II)/EDTA produced a dose-dependent increase in Ab F3/9 binding. This was reduced in the presence of catalase or aminoguanidine. We have validated the potential of gdC as a marker of oxidative DNA damage and showed negligible cross-reactivity with 8-oxo-2'-deoxyguanosine or malondialdehyde-modified DNA as well as its utility in immunocytochemistry. Formation of the gdC adduct may involve intermediate structures; however, our results strongly suggest Ab F3/9 has major specificity for the predominant product, 5-hydroxyacetyl-dC.

Cross-sectional studies in prescribing research

A cross-sectional study aims to describe the overall picture of a phenomenon, a situational problem, an attitude or an issue, by asking a cross-section of a given population at one specified moment in time. This paper describes the key features of the cross-sectional survey method. It begins by highlighting the main principles of the method, then discusses stages in the research process, drawing on two surveys of primary care pharmacists to illustrate some salient points about planning, sampling frames, definition and conceptual issues, research instrument design and response rates. Four constraints in prescribing studies were noted. First the newness of the subject meant a low basis of existing knowledge to design a questionnaire. Second, there was no public existing database for the sampling frame, so a pragmatic sampling exercise was used. Third, the definition of a Primary Care Pharmacist (PCP) [in full] and respondents recognition of that name and identification with the new role limited the response. Fourth, a growing problem for all surveys, but particularly with pharmacists and general practitioners (GP) [in full] is the growing danger of survey fatigue, which has a negative impact on response levels.

A developmental dissociation of view-dependent and view-invariant object recognition in adolescence

Spatial generalization skills in school children aged 8-16 were studied with regard to unfamiliar objects that had been previously learned in a cross-modal priming and learning paradigm. We observed a developmental dissociation with younger children recognizing objects only from previously learnt perspectives whereas older children generalized acquired object knowledge to new viewpoints as well. Haptic and - to a lesser extent - visual priming improved spatial generalization in all but the youngest children. The data supports the idea of dissociable, view-dependent and view-invariant object representations with different developmental trajectories that are subject to modulatory effects of priming. Late-developing areas in the parietal or the prefrontal cortex may account for the retarded onset of view-invariant object recognition. © 2006 Elsevier B.V. All rights reserved.

Towards a unified understanding of event-related changes in the EEG:the Firefly model of synchronization through cross-frequency phase modulation

Although event-related potentials (ERPs) are widely used to study sensory, perceptual and cognitive processes, it remains unknown whether they are phase-locked signals superimposed upon the ongoing electroencephalogram (EEG) or result from phase-alignment of the EEG. Previous attempts to discriminate between these hypotheses have been unsuccessful but here a new test is presented based on the prediction that ERPs generated by phase-alignment will be associated with event-related changes in frequency whereas evoked-ERPs will not. Using empirical mode decomposition (EMD), which allows measurement of narrow-band changes in the EEG without predefining frequency bands, evidence was found for transient frequency slowing in recognition memory ERPs but not in simulated data derived from the evoked model. Furthermore, the timing of phase-alignment was frequency dependent with the earliest alignment occurring at high frequencies. Based on these findings, the Firefly model was developed, which proposes that both evoked and induced power changes derive from frequency-dependent phase-alignment of the ongoing EEG. Simulated data derived from the Firefly model provided a close match with empirical data and the model was able to account for i) the shape and timing of ERPs at different scalp sites, ii) the event-related desynchronization in alpha and synchronization in theta, and iii) changes in the power density spectrum from the pre-stimulus baseline to the post-stimulus period. The Firefly Model, therefore, provides not only a unifying account of event-related changes in the EEG but also a possible mechanism for cross-frequency information processing.

Innate recognition of apoptotic cells:novel apoptotic cell-associated molecular patterns revealed by crossreactivity of anti-LPS antibodies

Cells dying by apoptosis are normally cleared by phagocytes through mechanisms that can suppress inflammation and immunity. Molecules of the innate immune system, the pattern recognition receptors (PRRs), are able to interact not only with conserved structures on microbes (pathogen-associated molecular patterns, PAMPs) but also with ligands displayed by apoptotic cells. We reasoned that PRRs might therefore interact with structures on apoptotic cells-apoptotic cell-associated molecular patterns (ACAMPs)-that are analogous to PAMPs. Here we show that certain monoclonal antibodies raised against the prototypic PAMP, lipopolysaccharide (LPS), can crossreact with apoptotic cells. We demonstrate that one such antibody interacts with a constitutively expressed intracellular protein, laminin-binding protein, which translocates to the cell surface during apoptosis and can interact with cells expressing the prototypic PRR, mCD14 as well as with CD14-negative cells. Anti-LPS cross reactive epitopes on apoptotic cells colocalised with annexin V-and C1q-binding sites on vesicular regions of apoptotic cell surfaces and were released associated with apoptotic cell-derived microvesicles (MVs). These results confirm that apoptotic cells and microbes can interact with the immune system through common elements and suggest that anti-PAMP antibodies could be used strategically to characterise novel ACAMPs associated not only with apoptotic cells but also with derived MVs. © 2013 Macmillan Publishers Limited All rights reserved.

Development of configural 3D object recognition

There is evidence for the late development in humans of configural face and animal recognition. We show that the recognition of artificial three-dimensional (3D) objects from part configurations develops similarly late. We also demonstrate that the cross-modal integration of object information reinforces the development of configural recognition more than the intra-modal integration does. Multimodal object representations in the brain may therefore play a role in configural object recognition. © 2003 Elsevier B.V. All rights reserved.

Estructura formal, textual y oral del discurso público

Estructura formal, textual y oral del discurso públicoExisten tres competencias comunicativas muy valoradas en la sociedad de la información. Primero, la búsqueda, selección y gestión de grandes cantidades de información. Segundo, la redacción de textos claros, concisos y rigurosos. Y en tercer lugar, la exposición y defensa oral de esta información en un discurso público. Tradicionalmente, los estudios de periodismo han abordado estas competencias de forma independiente. Pero actualmente, instituciones y empresas de ámbitos diferentes demandan un perfil profesional capaz de aplicarlas ante cualquier tipo de información y con objetivos diversos. Se propone un modelo integral en tres niveles estructurales basado en teorías, conceptos y estudios específicos de periodismo, oratoria, retórica… o comunicación, en los últimos años. Este modelo puede contribuir a encauzar las investigaciones de académicos y representa una herramienta de entrenamiento para profesionales.

Robust indoor speaker recognition in a network of audio and video sensors

Situational awareness is achieved naturally by the human senses of sight and hearing in combination. Automatic scene understanding aims at replicating this human ability using microphones and cameras in cooperation. In this paper, audio and video signals are fused and integrated at different levels of semantic abstractions. We detect and track a speaker who is relatively unconstrained, i.e., free to move indoors within an area larger than the comparable reported work, which is usually limited to round table meetings. The system is relatively simple: consisting of just 4 microphone pairs and a single camera. Results show that the overall multimodal tracker is more reliable than single modality systems, tolerating large occlusions and cross-talk. System evaluation is performed on both single and multi-modality tracking. The performance improvement given by the audio–video integration and fusion is quantified in terms of tracking precision and accuracy as well as speaker diarisation error rate and precision–recall (recognition). Improvements vs. the closest works are evaluated: 56% sound source localisation computational cost over an audio only system, 8% speaker diarisation error rate over an audio only speaker recognition unit and 36% on the precision–recall metric over an audio–video dominant speaker recognition method.

Pathways to emotional closeness in neonatal units - a cross-national qualitative study.

BACKGROUND: Research shows evidence for the importance of physical and emotional closeness for the infant, the parent and the infant-parent dyad. Less is known about how, when and why parents experience emotional closeness to their infants in a neonatal unit (NU), which was the aim of this study. METHODS: A qualitative study using a salutogenic approach to focus on positive health and wellbeing was undertaken in three NUs: one in Sweden, England and Finland. An 'emotional closeness' form was devised, which asked parents to describe moments/situations when, how and why they had felt emotionally close to their infant. Data for 23 parents of preterm infants were analyzed using thematic networks analysis. RESULTS: A global theme of 'pathways for emotional closeness' emerged from the data set. This concept related to how emotional, physical, cognitive and social influences led to feelings of emotional closeness between parents and their infants. The five underpinning organising themes relate to the: Embodied recognition through the power of physical closeness; Reassurance of, and contributing to, infant wellness; Understanding the present and the past; Feeling engaged in the day to day and Spending time and bonding as a family. CONCLUSION: These findings generate important insights into why, how and when parents feel emotionally close. This knowledge contributes to an increased awareness of how to support parents of premature infants to form positive and loving relationships with their infants. Health care staff should create a climate where parents' emotions and their emotional journey are individually supported.

Avoiding staff removal stage in optical music recognition: application to scores written in white mensural notation

Staff detection and removal is one of the most important issues in optical music recognition (OMR) tasks since common approaches for symbol detection and classification are based on this process. Due to its complexity, staff detection and removal is often inaccurate, leading to a great number of errors in posterior stages. For this reason, a new approach that avoids this stage is proposed in this paper, which is expected to overcome these drawbacks. Our approach is put into practice in a case of study focused on scores written in white mensural notation. Symbol detection is performed by using the vertical projection of the staves. The cross-correlation operator for template matching is used at the classification stage. The goodness of our proposal is shown in an experiment in which our proposal attains an extraction rate of 96 % and a classification rate of 92 %, on average. The results found have reinforced the idea of pursuing a new research line in OMR systems without the need of the removal of staff lines.

Delays In Receiving Obstetric Care And Poor Maternal Outcomes: Results From A National Multicentre Cross-sectional Study.

The vast majority of maternal deaths in low-and middle-income countries are preventable. Delay in obtaining access to appropriate health care is a fairly common problem which can be improved. The objective of this study was to explore the association between delay in providing obstetric health care and severe maternal morbidity/death. This was a multicentre cross-sectional study, involving 27 referral obstetric facilities in all Brazilian regions between 2009 and 2010. All women admitted to the hospital with a pregnancy-related cause were screened, searching for potentially life-threatening conditions (PLTC), maternal death (MD) and maternal near-miss (MNM) cases, according to the WHO criteria. Data on delays were collected by medical chart review and interview with the medical staff. The prevalence of the three different types of delays was estimated according to the level of care and outcome of the complication. For factors associated with any delay, the PR and 95%CI controlled for cluster design were estimated. A total of 82,144 live births were screened, with 9,555 PLTC, MNM or MD cases prospectively identified. Overall, any type of delay was observed in 53.8% of cases; delay related to user factors was observed in 10.2%, 34.6% of delays were related to health service accessibility and 25.7% were related to quality of medical care. The occurrence of any delay was associated with increasing severity of maternal outcome: 52% in PLTC, 68.4% in MNM and 84.1% in MD. Although this was not a population-based study and the results could not be generalized, there was a very clear and significant association between frequency of delay and severity of outcome, suggesting that timely and proper management are related to survival.