Dimensionality hyper-reduction and machine learning for dynamical systems with varying parameters

Thesis (Ph.D.)--University of Washington, 2016-08

The Role of G Protein In Alcohol Related Behaviours Using Drosophila melanogaster As A Model

Ethanol, classified as a drug, affects the central nervous system, and its consumption has been linked to the development of several behaviours including tolerance and dependence. Alcohol tolerance is defined as the need for higher doses of alcohol to induce the same changes observed in the initial exposure or where repetitive exposures of the same alcohol dose induce a lower response. Ethanol has been shown to interact with numerous targets and ultimately influence both short and long term adaptation at the cellular and molecular level in brain [1]. These adaptation processes are likely to involve signalling molecules: our work has focussed on G proteins gene expression. Using both wild type and several mutant fruit fly (Drosophila melanogaster) as a model for behaviour and molecular studies, we observed significant increases in sedation time (ST50) in response to alcohol (P<0.001) Fig.A. We also observed a consistent and significant decrease of Gq protein mRNA expression in Drosophila dUNC and DopR2 mutants chronically exposed to alcohol (*P<0.05). Fig B. Method: Six male flies were observed in drosophila polystyrene 25 x 95mm transparent vial in between cotton plugs. To the top plug, 500uL of 100% ethanol was added. Time till 50% of the flies were sedated was recorded on each day following the schedule. Fig. C (n=4-6). Using RT-PCR, we also quantified G protein mRNA expression levels one hour post initial 30 minutes of ethanol expression on day 1 and day 3 relative to expression in naïve flies.(n=2) [A] Increase in sedation time indicative of tolerance in different mutant lines and wild type flies. Six male flies were used in each experiment and (n= 4-6. ***P<0.001 unpaired t tests). [B] RT-PCR results showing significant reduction in Gq mRNA in flies chronically exposed to alcohol. (n=2. *P<0.05) [C] Alcohol exposure schedule. (1) Kaun K.R., R. Azanchi, Z. Maung, J. Hirsh, U. Heberlein. (2011). A Drosophila model for alcohol reward. Nature Neuroscience. 14 (5), 612–619.

The Effects of Systematic Self-Monitoring, Feedback, and Collaborative Assessment on Stress Reduction

Thesis (Ph.D.)--University of Washington, 2016-08

Development of a toolbox for the reduction of hordeins in barley malt beers

Gluten sensitive consumers and people suffering from coeliac disease account for up to 6% of the general population (Catassi et al., 2013). These consumers must avoid foods which contain gluten and related proteins found in wheat, rye or barley. Beer is produced from barley malt and therefore contains hordeins, (gluten like proteins). Beers labelled as gluten-free must contain below 10 mg/kg hordeins (10 mg/kg hordeins = 20 mg/kg gluten under current regulations) to be considered safe for gluten sensitive consumers. Currently there are a limited number of methods available for reducing beer hordeins, the studies outlined in this thesis provide a range of tools for the beverage industry to reduce the hordein content of beer It is well known, that during malting and brewing hordeins are reduced, but they still remain in beer at levels above 10 mg/kg. During malting, hordeins are broken down to form new proteins in the growing plant. Model malting and brewing systems were developed and used to test, how the modification of the malting process could be used to reduce beer hordeins. It was shown, that by using a controlled malting and brewing regime, a range of barley cultivars produced beer with significant differences in levels of hordeins. Beer hordeins ranged from 10 mg/kg to 60 mg/kg. Another study revealed that when malting was prolonged, to maximise breakdown of proteins, beer hordeins can be reduced by up to 44%. The natural breakdown of hordein during malting enhanced in a further study, when a protease was added to support the hordein degradation during steeping and germination. The enzyme addition resulted in a 46% reduction in beer hordeins 2 when compared to the control. All of the malt treatments had little or no impact on malt quality. The hordein levels can also be reduced during the beer stabilisation process. Levels of beer hordein were tested after stabilisation using two different concentrations of silica gel and tannic acid. Silica gel was very effective in reducing beer hordeins, 90% of beer hordeins were removed compared to the control beer. Beer hordeins could be reduced to below 10 mg/kg and the beer qualities such as foam, colour and flavour were not affected. Tannic acid also reduced beer hordein by up to 90%, but it reduced foam stability and affected beer flavours. A further study described treatment of beer with microbial transglutaminase (mTG), to create bonds between hordein proteins, which increased particle size and allowed removal during filtration. The addition of the mTG led to a reduction of the beer hordein by up to 96% in beer, and the impact on the resulting beer quality was minimal. These studies provide the industry with a toolbox of methods leading to the reduction of hordein in the final beer without negatively affecting beer quality.

Consumer perceived risk, risk reduction strategies and transaction intentions in online marketplaces

Even though online commerce has garnered vast academic interest during the recent years, theoretical grounds for consumer behavior online still remains ambiguous. Despite the globally rapid growth of online commerce, only a fraction of Internet browsers end up purchasing goods online. This is argued to be caused by the intangible and distant nature of the Internet, causing overwhelming perceived risks for consumers and negatively affecting transaction intentions. To combat perceived risks, consumers may actively or passively seek to relieve those risks to tolerable level. These risk reduction strategies refer to both institutional mechanisms as well as consumer risk reduction strategies. The objective of this thesis is to provide further understanding upon the relationships between consumer perceived risk, risk reduction strategies and transaction intentions in online marketplaces. To serve the objectives of the present thesis, a quantitative approach was chosen as the method for conducting empirical research. The data was collected with an online survey through discussion board, using a random sample approach. The proposed research model was examined with a set of hierarchical regression analyses. Results revealed several direct relationships as well as moderating interaction effects. The key finding of this thesis is that institutional risk reduction mechanisms significantly contribute to consumer perceived risks. These mechanisms have the potential to reduce perceived risks, and therefore may stimulate transaction intentions. Additionally, it was observed that risk reduction strategies moderate the relationship between intermediary provided risk relievers, consumer perceived risks and transaction intentions. Retailer related risk reduction strategies were also shown to enforce the effectiveness of payment methods; however feedback and monitoring mechanism was shown to have a diminishing effect of perceived risk only when consumers did not rely on product related risk reduction strategies. The present thesis also illustrates the importance of effective information search, as those consumers are more willing to transact as the perceived risks become less significant. For managerial purposes, the importance of well-functioning institutional mechanisms cannot be emphasized enough.

Multi-physics modeling of aluminium reduction cells

Aluminium cells involve a range of complex physical processes which act simultaneously to provide a narrow satisfactory operating range. These processes involve electromagnetic fields, coupled with heat transfer and phase change, two phase fluid flow with a range of complexities plus the development of stress in the cell structure. All of these phenomena are coupled in some significant sense and so to provide a comprehensive model of these processes involves their representation simultaneously. Conventionally, aspects of the process have been modeled separately using uncoupled estimates of the effects of the other phenomena; this has enabled the use of standard commercial CFD and FEA tools. In this paper we will describe an approach to the modeling of aluminium cells which describes all the physics simultaneously. This approach uses a finite volume approximation for each of the phenomena and facilitates their interactions directly in the modeling-the complex geometries involved are addressed by using unstructured meshes. The very challenging issues to be overcome in this venture will be outlined and some preliminary results will be shown.

Phase noise and spur reduction in an array of direct digital synthesizers

Many applications, including communications, test and measurement, and radar, require the generation of signals with a high degree of spectral purity. One method for producing tunable, low-noise source signals is to combine the outputs of multiple direct digital synthesizers (DDSs) arranged in a parallel configuration. In such an approach, if all noise is uncorrelated across channels, the noise will decrease relative to the combined signal power, resulting in a reduction of sideband noise and an increase in SNR. However, in any real array, the broadband noise and spurious components will be correlated to some degree, limiting the gains achieved by parallelization. This thesis examines the potential performance benefits that may arise from using an array of DDSs, with a focus on several types of common DDS errors, including phase noise, phase truncation spurs, quantization noise spurs, and quantizer nonlinearity spurs. Measurements to determine the level of correlation among DDS channels were made on a custom 14-channel DDS testbed. The investigation of the phase noise of a DDS array indicates that the contribution to the phase noise from the DACs can be decreased to a desired level by using a large enough number of channels. In such a system, the phase noise qualities of the source clock and the system cost and complexity will be the main limitations on the phase noise of the DDS array. The study of phase truncation spurs suggests that, at least in our system, the phase truncation spurs are uncorrelated, contrary to the theoretical prediction. We believe this decorrelation is due to the existence of an unidentified mechanism in our DDS array that is unaccounted for in our current operational DDS model. This mechanism, likely due to some timing element in the FPGA, causes some randomness in the relative phases of the truncation spurs from channel to channel each time the DDS array is powered up. This randomness decorrelates the phase truncation spurs, opening the potential for SFDR gain from using a DDS array. The analysis of the correlation of quantization noise spurs in an array of DDSs shows that the total quantization noise power of each DDS channel is uncorrelated for nearly all values of DAC output bits. This suggests that a near N gain in SQNR is possible for an N-channel array of DDSs. This gain will be most apparent for low-bit DACs in which quantization noise is notably higher than the thermal noise contribution. Lastly, the measurements of the correlation of quantizer nonlinearity spurs demonstrate that the second and third harmonics are highly correlated across channels for all frequencies tested. This means that there is no benefit to using an array of DDSs for the problems of in-band quantizer nonlinearities. As a result, alternate methods of harmonic spur management must be employed.

The tetrahexahedric angular Calogero model

The spherical reduction of the rational Calogero model (of type A n−1 and after removing the center of mass) is considered as a maximally superintegrable quantum system, which describes a particle on the (n−2)-sphere subject to a very particular potential. We present a detailed analysis of the simplest non-separable case, n=4, whose potential is singular at the edges of a spherical tetrahexahedron. A complete set of independent conserved charges and of Hamiltonian intertwiners is constructed, and their algebra is elucidated. They arise from the ring of polynomials in Dunkl-deformed angular momenta, by classifying the subspaces invariant and antiinvariant under all Weyl reflections, respectively.

Data Reduction in Intrusion Alert Correlation

Network intrusion detection sensors are usually built around low level models of network traffic. This means that their output is of a similarly low level and as a consequence, is difficult to analyze. Intrusion alert correlation is the task of automating some of this analysis by grouping related alerts together. Attack graphs provide an intuitive model for such analysis. Unfortunately alert flooding attacks can still cause a loss of service on sensors, and when performing attack graph correlation, there can be a large number of extraneous alerts included in the output graph. This obscures the fine structure of genuine attacks and makes them more difficult for human operators to discern. This paper explores modified correlation algorithms which attempt to minimize the impact of this attack.

The Role of Cannabinoids in the Neurobiology of Sensory Gating: A firing rate model study

Gating of sensory (e.g. auditory) information has been demonstrated as a reduction in the auditory-evoked potential responses recorded in the brain of both normal animals and human subjects. Auditory gating is perturbed in schizophrenic patients and pharmacologically by drugs such as amphetamine, phencyclidine or ketamine, which precipitate schizophrenic-like symptoms in normal subjects. The neurobiological basis underlying this sensory gating can be investigated using local field potential recordings from single electrodes. In this paper we use such technology to investigate the role of cannabinoids in sensory gating. Cannabinoids represent a fundamentally new class of retrograde messengers which are released postsynaptically and bind to presynaptic receptors. In this way they allow fine-tuning of neuronal response, and in particular can lead to so-called depolarization-induced suppression of inhibition (DSI). Our experimental results show that application of the exogenous cannabinoid WIN55, 212-2 can abolish sensory gating as measured by the amplitude of local field responses in rat hippocampal region CA3. Importantly we develop a simple firing rate population model of CA3 and show that gating is heavily dependent upon the presence of a slow inhibitory (GABAB) pathway. Moreover, a simple phenomenological model of cannabinoid dynamics underlying DSI is shown to abolish gating in a manner consistent with our experimental findings.

Down scaling to regional assessment of greenhouse gas emissions to enable consistency in accounting for emissions reduction projects and national inventory accounts for northern beef production in Australia

This paper explores the effect of using regional data for livestock attributes on estimation of greenhouse gas (GHG) emissions for the northern beef industry in Australia, compared with using state/territory-wide values, as currently used in Australia’s national GHG inventory report. Regional GHG emissions associated with beef production are reported for 21 defined agricultural statistical regions within state/territory jurisdictions. A management scenario for reduced emissions that could qualify as an Emissions Reduction Fund (ERF) project was used to illustrate the effect of regional level model parameters on estimated abatement levels. Using regional parameters, instead of state level parameters, for liveweight (LW), LW gain and proportion of cows lactating and an expanded number of livestock classes, gives a 5.2% reduction in estimated emissions (range +12% to –34% across regions). Estimated GHG emissions intensity (emissions per kilogram of LW sold) varied across the regions by up to 2.5-fold, ranging from 10.5 kg CO2-e kg–1 LW sold for Darling Downs, Queensland, through to 25.8 kg CO2-e kg–1 LW sold for the Pindan and North Kimberley, Western Australia. This range was driven by differences in production efficiency, reproduction rate, growth rate and survival. This suggests that some regions in northern Australia are likely to have substantial opportunities for GHG abatement and higher livestock income. However, this must be coupled with the availability of management activities that can be implemented to improve production efficiency; wet season phosphorus (P) supplementation being one such practice. An ERF case study comparison showed that P supplementation of a typical-sized herd produced an estimated reduction of 622 t CO2-e year–1, or 7%, compared with a non-P supplemented herd. However, the different model parameters used by the National Inventory Report and ERF project means that there was an anomaly between the herd emissions for project cattle excised from the national accounts (13 479 t CO2-e year–1) and the baseline herd emissions estimated for the ERF project (8 896 t CO2-e year–1) before P supplementation was implemented. Regionalising livestock model parameters in both ERF projects and the national accounts offers the attraction of being able to more easily and accurately reflect emissions savings from this type of emissions reduction project in Australia’s national GHG accounts.

Vocabulary and language model adaptation using information retrieval

International audience

Implementation of the CDM model to the analisys of the hotel sector

The main aim of this study was to determine the impact of innovation on productivity in service sector companies — especially those in the hospitality sector — that value the reduction of environmental impact as relevant to the innovation process. We used a structural analysis model based on the one developed by Crépon, Duguet, and Mairesse (1998). This model is known as the CDM model (an acronym of the authors’ surnames). These authors developed seminal studies in the field of the relationships between innovation and productivity (see Griliches 1979; Pakes and Grilliches 1980). The main advantage of the CDM model is its ability to integrate the process of innovation and business productivity from an empirical perspective.

Exogenous Administration of 15d-PGJ(2)-Loaded Nanocapsules Inhibits Bone Resorption in a Mouse Periodontitis Model

The 15-deoxy-(Delta 12,14)-PG J(2) (15d-PGJ(2)) has demonstrated excellent anti-inflammatory results in different experimental models. It can be used with a polymeric nanostructure system for modified drug release, which can change the therapeutic properties of the active principle, leading to increased stability and slower/prolonged release. The aim of the current study was to test a nano-technological formulation as a carrier for 15d-PGJ(2), and to investigate the immunomodulatory effects of this formulation in a mouse periodontitis model. Poly (D, L-lactide-coglycolide) nanocapsules (NC) were used to encapsulate 15d-PGJ(2). BALB/c mice were infected on days 0, 2, and 4 with Aggregatibacter actinomycetemcomitans and divided into groups (n = 5) that were treated daily during 15 d with 1, 3, or 10 mu g/kg 15d-PGJ(2)-NC. The animals were sacrificed, the submandibular lymph nodes were removed for FACS analysis, and the jaws were analyzed for bone resorption by morphometry. Immunoinflammatory markers in the gingival tissue were analyzed by reverse transcriptase-quantitative PCR, Western blotting, or ELISA. Infected animals treated with the 15d-PGJ(2)-NC presented lower bone resorption than infected animals without treatment (p < 0.05). Furthermore, infected animals treated with 10 mu g/kg 15d-PGJ(2)-NC had a reduction of CD4(+)CD25(+)FOXP3(+) cells and CD4/CD8 ratio in the submandibular lymph node (p < 0.05). Moreover, CD55 was upregulated, whereas RANKL was downregulated in the gingival tissue of the 10 mu g/kg treated group (p < 0.05). Several proinflammatory cytokines were decreased in the group treated with 10 mu g/kg 15d-PGJ(2)-NC, and high amounts of 15d-PGJ(2) were observed in the gingiva. In conclusion, the 15d-PGJ(2)-NC formulation presented immunomodulatory effects, decreasing bone resorption and inflammatory responses in a periodontitis mouse model. The Journal of Immunology, 2012, 189: 1043-1052.

A Learning Function for Parameter Reduction in Spiking Neural Networks with Radial Basis Function

Spiking neural networks - networks that encode information in the timing of spikes - are arising as a new approach in the artificial neural networks paradigm, emergent from cognitive science. One of these new models is the pulsed neural network with radial basis function, a network able to store information in the axonal propagation delay of neurons. Learning algorithms have been proposed to this model looking for mapping input pulses into output pulses. Recently, a new method was proposed to encode constant data into a temporal sequence of spikes, stimulating deeper studies in order to establish abilities and frontiers of this new approach. However, a well known problem of this kind of network is the high number of free parameters - more that 15 - to be properly configured or tuned in order to allow network convergence. This work presents for the first time a new learning function for this network training that allow the automatic configuration of one of the key network parameters: the synaptic weight decreasing factor.