abstract: Mixture of experts is a machine learning ensemble approach that consists of individual models that are trained to be ``experts'' on subsets of the data, and a gating network that provides weights to output a combination of the expert predictions. Mixture of experts models do not currently see wide use due to difficulty in training diverse experts and high computational requirements. This work presents modifications of the mixture of experts formulation that use domain knowledge to improve training, and incorporate parameter sharing among experts to reduce computational requirements. First, this work presents an application of mixture of experts models for quality robust visual recognition. First it is shown that human subjects outperform deep neural networks on classification of distorted images, and then propose a model, MixQualNet, that is more robust to distortions. The proposed model consists of ``experts'' that are trained on a particular type of image distortion. The final output of the model is a weighted sum of the expert models, where the weights are determined by a separate gating network. The proposed model also incorporates weight sharing to reduce the number of parameters, as well as increase performance. Second, an application of mixture of experts to predict visual saliency is presented. A computational saliency model attempts to predict where humans will look in an image. In the proposed model, each expert network is trained to predict saliency for a set of closely related images. The final saliency map is computed as a weighted mixture of the expert networks' outputs, with weights determined by a separate gating network. The proposed model achieves better performance than several other visual saliency models and a baseline non-mixture model. Finally, this work introduces a saliency model that is a weighted mixture of models trained for different levels of saliency. Levels of saliency include high saliency, which corresponds to regions where almost all subjects look, and low saliency, which corresponds to regions where some, but not all subjects look. The weighted mixture shows improved performance compared with baseline models because of the diversity of the individual model predictions. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2018

Evidence suggests that the gut-brain axis can influence stress-related behaviour, mood and neuropsychological disorders, including anxiety. Stress exposure can increase anxiety-related symptoms such as muscle tension & worrying. Medical treatment has low success and a range of side effects on anxiety. This review aimed to see if probiotics can reduce anxiety symptoms in humans. Where relevant articles on people with anxiety disorders are lacking, the review evaluates articles addressing healthy participants in stressful situations such as exams or public speeches via anxiety questionnaires. I hypothesized that probiotics could be an effective anxiolytic treatment in combination with therapy. Most articles demonstrated reduced subjective and objective results in anxiety and stress measurements after a daily intake of probioticstrains. Findings demonstrate potential anxiolytic benefits with a daily probiotic intake. However, future research on participants with an anxiety disorder is needed to conclude the hypothesis.

In addition to fetal alcohol syndrome and fetal alcohol spectrum disorders, prenatal alcohol exposure is associated with many other adverse pregnancy and birth outcomes. Research suggests that alcohol use during pregnancy may increase the risk of miscarriage, stillbirth, preterm delivery, and sudden infant death syndrome. This research has some inherent difficulties, such as the collection of accurate information about alcohol consumption during pregnancy and controlling for comorbid exposures and conditions. Consequently, attributing poor birth outcomes to prenatal alcohol exposure is a complicated and ongoing task, requiring continued attention to validated methodology and to identifying specific biological mechanisms.