Detection and Classification of Whale Acoustic Signals

This dissertation focuses on two vital challenges in relation to whale acoustic signals: detection and classification.

In detection, we evaluated the influence of the uncertain ocean environment on the spectrogram-based detector, and derived the likelihood ratio of the proposed Short Time Fourier Transform detector. Experimental results showed that the proposed detector outperforms detectors based on the spectrogram. The proposed detector is more sensitive to environmental changes because it includes phase information.

In classification, our focus is on finding a robust and sparse representation of whale vocalizations. Because whale vocalizations can be modeled as polynomial phase signals, we can represent the whale calls by their polynomial phase coefficients. In this dissertation, we used the Weyl transform to capture chirp rate information, and used a two dimensional feature set to represent whale vocalizations globally. Experimental results showed that our Weyl feature set outperforms chirplet coefficients and MFCC (Mel Frequency Cepstral Coefficients) when applied to our collected data.

Since whale vocalizations can be represented by polynomial phase coefficients, it is plausible that the signals lie on a manifold parameterized by these coefficients. We also studied the intrinsic structure of high dimensional whale data by exploiting its geometry. Experimental results showed that nonlinear mappings such as Laplacian Eigenmap and ISOMAP outperform linear mappings such as PCA and MDS, suggesting that the whale acoustic data is nonlinear.

We also explored deep learning algorithms on whale acoustic data. We built each layer as convolutions with either a PCA filter bank (PCANet) or a DCT filter bank (DCTNet). With the DCT filter bank, each layer has different a time-frequency scale representation, and from this, one can extract different physical information. Experimental results showed that our PCANet and DCTNet achieve high classification rate on the whale vocalization data set. The word error rate of the DCTNet feature is similar to the MFSC in speech recognition tasks, suggesting that the convolutional network is able to reveal acoustic content of speech signals.

Ink Under the Fingernails: Making Print in Nineteenth-Century Mexico City

This dissertation examines Mexico City’s material politics of print—the central actors engaged in making print, their activities and relationships, and the legal, business, and social dimensions of production—across the nineteenth century. Inside urban printshops, a socially diverse group of men ranging from manual laborers to educated editors collaborated to make the printed items that fueled political debates and partisan struggles in the new republic. By investigating how print was produced, regulated, and consumed, this dissertation argues that printers shaped some of the most pressing conflicts that marked Mexico’s first formative century: over freedom of expression, the role of religion in government, and the emergence of liberalism. Printers shaped debates not only because they issued texts that fueled elite politics but precisely because they operated at the nexus where new liberal guarantees like freedom of the press and intellectual property intersected with politics and patronage, the regulatory efforts of the emerging state, and the harsh realities of a post-colonial economy.

Historians of Mexico have typically approached print as a vehicle for texts written by elites, which they argue contributed to the development of a national public sphere or print culture in spite of low literacy levels. By shifting the focus to print’s production, my work instead reveals that a range of urban residents—from prominent printshop owners to government ministers to street vendors—produced, engaged, and deployed printed items in contests unfolding in the urban environment. As print increasingly functioned as a political weapon in the decades after independence, print production itself became an arena in struggles over the emerging contours of politics and state formation, even as printing technologies remained relatively unchanged over time.

This work examines previously unexplored archival documents, including official correspondence, legal cases, business transactions, and printshop labor records, to shed new light on Mexico City printers’ interactions with the emerging national government, and reveal the degree to which heated ideological debates emerged intertwined with the most basic concerns over the tangible practices of print. By delving into the rich social and cultural world of printing—described by intellectuals and workers alike in memoirs, fiction, caricatures and periodicals— it also considers how printers’ particular status straddling elite and working worlds led them to challenge boundaries drawn by elites that separated manual and intellectual labors. Finally, this study engages the full range of printed documents made in Mexico City printshops not just as texts but also as objects with particular visual and material qualities whose uses and meanings were shaped not only by emergent republicanism but also by powerful colonial legacies that generated ambivalent attitudes towards print’s transformative power.

The Political Economy of Climate Adaptation and Environmental Health: The Case of Ethiopia

The environment affects our health, livelihoods, and the social and political institutions within which we interact. Indeed, nearly a quarter of the global disease burden is attributed to environmental factors, and many of these factors are exacerbated by global climate change. Thus, the central research question of this dissertation is: How do people cope with and adapt to uncertainty, complexity, and change of environmental and health conditions? Specifically, I ask how institutional factors, risk aversion, and behaviors affect environmental health outcomes. I further assess the role of social capital in climate adaptation, and specifically compare individual and collective adaptation. I then analyze how policy develops accounting for both adaptation to the effects of climate and mitigation of climate-changing emissions. In order to empirically test the relationships between these variables at multiple levels, I combine multiple methods, including semi-structured interviews, surveys, and field experiments, along with health and water quality data. This dissertation uses the case of Ethiopia, Africa’s second-most populous nation, which has a large rural population and is considered very vulnerable to climate change. My fieldwork included interviews and institutional data collection at the national level, and a three-year study (2012-2014) of approximately 400 households in 20 villages in the Ethiopian Rift Valley. I evaluate the theoretical relationships between households, communities, and government in the process of adaptation to environmental stresses. Through my analyses, I demonstrate that water source choice varies by individual risk aversion and institutional context, which ultimately has implications for environmental health outcomes. I show that qualitative measures of trust predict cooperation in adaptation, consistent with social capital theory, but that measures of trust are negatively related with private adaptation by the individual. Finally, I describe how Ethiopia had some unique characteristics, significantly reinforced by international actors, that led to the development of an extensive climate policy, and yet with some challenges remaining for implementation. These results suggest a potential for adaptation through the interactions among individuals, communities, and government in the search for transformative processes when confronting environmental threats and climate change.

Efficient Prices at Any Cost: Does Algorithmic Trading Deter Information Acquisition?

I demonstrate a powerful tension between acquiring information and incorporating it into asset prices, the two core elements of price discovery. As a salient case, I focus on the transformative rise of algorithmic trading (AT) typically associated with improved price efficiency. Using a measure of the relative information content of prices and a comprehensive panel of 37,325 stock-quarters of SEC market data, I establish instead that algorithmic trading strongly decreases the net amount of information in prices. The increase in price distortions associated with the AT “information gap” is roughly $42.6 billion/year for U.S. common stocks around earnings announcement events alone. Information losses are concentrated among stocks with high shares of algorithmic liquidity takers relative to algorithmic liquidity makers, suggesting that aggressive AT powerfully deters fundamental information acquisition despite its importance for translating available information into prices.

Inter-site and inter-scanner diffusion MRI data harmonization.

We propose a novel method to harmonize diffusion MRI data acquired from multiple sites and scanners, which is imperative for joint analysis of the data to significantly increase sample size and statistical power of neuroimaging studies. Our method incorporates the following main novelties: i) we take into account the scanner-dependent spatial variability of the diffusion signal in different parts of the brain; ii) our method is independent of compartmental modeling of diffusion (e.g., tensor, and intra/extra cellular compartments) and the acquired signal itself is corrected for scanner related differences; and iii) inter-subject variability as measured by the coefficient of variation is maintained at each site. We represent the signal in a basis of spherical harmonics and compute several rotation invariant spherical harmonic features to estimate a region and tissue specific linear mapping between the signal from different sites (and scanners). We validate our method on diffusion data acquired from seven different sites (including two GE, three Philips, and two Siemens scanners) on a group of age-matched healthy subjects. Since the extracted rotation invariant spherical harmonic features depend on the accuracy of the brain parcellation provided by Freesurfer, we propose a feature based refinement of the original parcellation such that it better characterizes the anatomy and provides robust linear mappings to harmonize the dMRI data. We demonstrate the efficacy of our method by statistically comparing diffusion measures such as fractional anisotropy, mean diffusivity and generalized fractional anisotropy across multiple sites before and after data harmonization. We also show results using tract-based spatial statistics before and after harmonization for independent validation of the proposed methodology. Our experimental results demonstrate that, for nearly identical acquisition protocol across sites, scanner-specific differences can be accurately removed using the proposed method.