Browsing by Subject "Canadian history"

A fire in Salem, Mass., in 1914 and an explosion in Halifax, N.S., in 1917 provide an opportunity to explore working-class institutions and organizations in the United States-Canada borderlands. In a historical moment in which the state greatly expanded its responsibility to give protection and rescue to its citizens, after these two disasters ordinary survivors preferred to depend on their friends, neighbors, and family members. This dissertation examines which institutions--including formal organizations like unions and fraternal societies as well as informal groups like families and neighborhoods--were most relevant and useful to working-class survivors. Families, neighbors, friends, and coworkers had patterns and traditions of self-help, informal order, and solidarity that they developed before crisis hit their cities. Those traditions were put to unusual purposes and extreme stress when the disasters happened. They were also challenged by new agents of the state, who were given extraordinary powers in the wake of the disasters. This dissertation describes how the working-class people who most directly experienced the disasters understood them and their cities starkly differently than the professionalized relief authorities.

Using a wide array of sources--including government documents, published accounts, archived ephemeral, oral histories, photographs, newspapers in two languages, and the case files of the Halifax Relief Commission--the dissertation describes how elites imposed a progressive state on what they imagined to be a fractured and chaotic social landscape. It argues that "the people" for whom reformers claimed to speak had their own durable, alternative modes of support and rescue that they quickly and effectively mobilized in times of crisis, but which remained illegible to elites. By demonstrating the personal, ideological, political, and practical ties between New England and Nova Scotia and Quebec, it also emphasizes the importance of studying American and Canadian history together, not only comparatively but as a transnational, North American whole.

Recently there has been increasing interest in developing generative models of data, offering the promise of learning based on the often vast quantity of unlabeled data. With such learning, one typically seeks to build rich, hierarchical probabilistic models that are able to

fit to the distribution of complex real data, and are also capable of realistic data synthesis. In this dissertation, novel models and learning algorithms are proposed for deep generative models.

This disseration consists of three main parts.

The first part developed a deep generative model joint analysis of images and associated labels or captions. The model is efficiently learned using variational autoencoder. A multilayered (deep) convolutional dictionary representation is employed as a decoder of the

latent image features. Stochastic unpooling is employed to link consecutive layers in the image model, yielding top-down image generation. A deep Convolutional Neural Network (CNN) is used as an image encoder; the CNN is used to approximate a distribution for the latent DGDN features/code. The latent code is also linked to generative models for labels (Bayesian support vector machine) or captions (recurrent neural network). When predicting a label/caption for a new image at test, averaging is performed across the distribution of latent codes; this is computationally efficient as a consequence of the learned CNN-based encoder. Since the framework is capable of modeling the image in the presence/absence of associated labels/captions, a new semi-supervised setting is manifested for CNN learning with images; the framework even allows unsupervised CNN learning, based on images alone. Excellent results are obtained on several benchmark datasets, including ImageNet, demonstrating that the proposed model achieves results that are highly competitive with similarly sized convolutional neural networks.

The second part developed a new method for learning variational autoencoders (VAEs), based on Stein variational gradient descent. A key advantage of this approach is that one need not make parametric assumptions about the form of the encoder distribution. Performance is further enhanced by integrating the proposed encoder with importance sampling. Excellent performance is demonstrated across multiple unsupervised and semi-supervised problems, including semi-supervised analysis of the ImageNet data, demonstrating the scalability of the model to large datasets.

The third part developed a new form of variational autoencoder, in which the joint distribution of data and codes is considered in two (symmetric) forms: (i) from observed data fed through the encoder to yield codes, and (ii) from latent codes drawn from a simple

prior and propagated through the decoder to manifest data. Lower bounds are learned for marginal log-likelihood fits observed data and latent codes. When learning with the variational bound, one seeks to minimize the symmetric Kullback-Leibler divergence of

joint density functions from (i) and (ii), while simultaneously seeking to maximize the two marginal log-likelihoods. To facilitate learning, a new form of adversarial training is developed. An extensive set of experiments is performed, in which we demonstrate state-of-the-art data reconstruction and generation on several image benchmark datasets.

Dreams of a "tropical Canada" that included the West Indies occupied the thoughts of many Canadians over a period spanning nearly forty years. From the expansionist fever of the late nineteenth century to the redistribution of German territories immediately following the First World War, Canadians of varying backgrounds campaigned vigorously for Canada-West Indies union. Their efforts generated a transatlantic discourse that raised larger questions about Canada's national trajectory, imperial organization, and the state of Britain's Empire in the twentieth century.

This dissertation explores the key ideas, tensions, and contradictions that shaped the union discourse over time. Race, nation and empire were central to this discourse. Canadian expansionists' efforts to gain free access to tropical territory, consolidate British possessions in the Western hemisphere, and negotiate the terms under which West Indians of color would enter the Canadian federation reflected and perpetuated logics that were simultaneously racial, national, and imperial.

Canada-West Indies union campaigns raise important questions about the processes at work in the ideological and material formation of the Canadian "nation" in the late nineteenth and early twentieth century. Employing a wide range of public and private manuscript material, diaries, travelogues and newspapers, this dissertation argues that Canadians' expansionist aspirations in the West Indies were inextricably connected to a national vision. To the campaign's advocates, acquiring colonial satellites - particularly in tropical regions - was a defining feature of nation-state formation.

This dissertation is comprised of three studies that investigate the implications and determinants of public funding for universities.

The first chapter lays down the foundation for the other two studies. I discuss how state-level policies, as determined by legislators, represent a pivotal component of firms' non-market strategies that have direct implications for the viability of their innovative and entrepreneurial activities. I expand this discussion to identify gaps in extant studies surrounding state-level policies and state legislators.

The second study focuses on the state funding for 420 public universities to estimate the precise return on state investments in higher education as measured by two economic outcomes -- the generation of university patents and formation of business establishments in a given university's local economy from 2002 through 2014. Using an instrumental variable estimation strategy, I predict and find a positive, causal association between state funding and the number of patents granted to public universities. I also observe a similar causal relationship between state funding and the entry of new business establishments near a given campus. This becomes pronounced for small firms in the manufacturing, retail, and service sectors, and even more for small firms in high-technology industries that are known to rely heavily on universities as a source of external inventions - pharmaceutical, medical equipment, and semiconductor.

The third study explores a new determinant of state funding for 420 public universities by leveraging novel, hand-collected data on the educational experiences of state legislators - specifically if and where they received postsecondary education. I predict and find a statistically significant, positive association between the share of legislators who attended their states' public institutions and state funding for their entire public higher-education system. A similar positive relationship is also observed between the share of state legislators who attended particular campuses of the state's public university system and funding for those campuses. This relationship is more pronounced among publicly educated legislators who represent legislative districts close to their alma mater's district, and becomes most consequential when the legislator's district contains his or her alma mater.