The promise of optical antennas is the ability to tame the light to behave in ways not achievable using traditional optical components. For example, our results here demonstrate that a careful engineering of optical antennas allow the strong, even perfect, absorption of light in ultra-thin geometries, i.e., geometries much thinner than the wavelength of light. Enabled by geometry-sensitive antenna resonances, this absorption behavior can also be realized for a broad selection of colors. A detailed theoretical analysis of the observed perfect absorption phenomenon reveals the role of incoherently interacting degenerate electric and magnetic resonances in overcoming the well-known absorption limit for infinitesimally thin films. With another set of experiments, we show that strongly absorbed optical energy in aluminum nanoantennas can be used to heat them efficiently above their melting temperature and stimulate an explosive exothermic oxidation reaction called melt-dispersion mechanism. Importantly, we see that engineering the specific geometry of the constituent particles allows an unprecedented control of aluminum ignition, both spectrally and spatially, through the fine tuning of the optical antenna resonances.

The technological revolution that started with digital electronics more than 50 years ago has pushed the limits of scalability in device fabrication. Device features are currently at the nanometer scale, which requires structures to be built with atomic level accuracy. Advances in nanotechnology have opened up an exciting area of research that allows for molecular level control of device surfaces. Organic molecules can provide the tailorability required to continue the progress of semiconductor technologies. Organic functionalization provides a pathway to control the surface at the molecular level. Fundamental understanding of the adsorption phenomena between organic molecules and the surface is critical to achieve a stable inorganic/organic interface for the creation of hybrid nanostructures. This thesis aims to expand our current toolkit on functionalization to molecules that have multiple functionalities that can react with the surface. The reaction mechanism of these molecules is complex as there are several driving forces that can play a role during adsorption and influence the final reaction products. This thesis covers the adsorption of multifunctional molecules on the Ge(100)-2×1 surface using a combination of experimental and theoretical techniques: Fourier transform infrared spectroscopy, X-ray photoelectron spectroscopy, and density functional theory calculations. We explored the adsorption of four different molecules: 1,2,3-benzenetriol (C6H6O3); 1,3,5-benzenetriol (C6H6O3); 2-hydroxymethyl-1,3-propanediol (C4H10O3); pyrazine (C4H6N2). These are the first reported studies of these molecules on the Ge(100) surface. In order to understand the extent to which molecular geometry affects surface coverage, a detailed comparison of the adsorption of two triol molecules was carried out: 1,3,5-benzenetriol which has a rigid phenyl backbone and 2-hydroxymethyl-1,3-propanediol with a flexible alkyl backbone. DFT results showed that the rigid backbone exhibits a higher degree of strain, which translates to a loss of exothermicity in the reaction coordinate. Experiments showed that the flexibility of the alkyl backbone provides higher rotational degrees of freedom and an enhancement in surface coverage. In order to understand the effect of intermolecular interactions, the adsorption of 1,2,3-benzenetriol was explored. Interestingly, we found that at high coverage, intramolecular hydrogen bonding in singly and dually adsorbate products breaks to form intermolecular hydrogen bonding with a nearby adsorbate, which provides enhanced stabilization of the surface adduct. This additional stabilization may lower the reactivity of unreacted functional groups even if an empty nearby Ge site is available for reaction. The distribution of products resulted in primarily bidentate adsorbates, leaving an unreacted moiety at the surface. We also found evidence of coverage and temperature effects on adsorbed pyrazine molecules on the Ge surface. It was observed that this molecule adsorbs on Ge through both carbon and nitrogen moieties and the product distribution changes as a function of coverage and temperature. At low coverage, incoming molecules react primarily through C-cycloaddition reactions and N dative bonds. However, as the density of adsorbates increases, new incoming molecules adsorbed primarily through the nitrogen moiety. Furthermore, as the temperature increases, the product distribution changed from primarily non-activated dative bond to activated cycloaddition products. Overall, the studies in this thesis provide new insight into competition and selectivity in adsorption of multifunctional molecules on the Ge(100)-2×1 surface.

Changes in regulation enacted in 2013 have enabled the Alaska Fisheries Science Center's Fishery Monitoring and Analysis Division (FMA) and the Alaska Regional Office's Sustainable Fisheries Division to work collaboratively on an Annual Deployment Plan (ADP). Each ADP documents how the National Marine Fisheries Service (NMFS) plans to deploy observers into fishing activities for the coming year under the limits of available funding. Draft ADPs are presented to the North Pacific Fishery Management Council (Council) during September - October and are finalized in December. The sampling design for observer deployment has two elements: how the population is subdivided (i.e., stratification schemes) and how available samples are allocated (i.e., allocation strategies). Here the relative performance of 10 alternative sampling designs (at the primary sampling unit- the trip) are compared in support of the draft 2018 ADP. These alternative sampling designs consisted of the combination of two stratification schemes (gear-type only or gear-type × tendering activity), two metrics upon which to base optimizations [one consisting of discard of groundfish with Prohibited Species Catch (PSC) of Pacific halibut and the other consisting of the prior and PSC of Chinook salmon], and three allocation strategies (no optimization, a "hurdle" approach to optimization, and a optimization only). All optimization allocations incorporate three variables measured over the past 3 years: variance in the metric, the average cost of observing a trip, and the number of trips. Total afforded sample size is determined by the available budget and the average cost of observing each trip. Resulting selection rates derive from sample size, allocation weightings and the anticipated fishing effort which was defined as the most recent complete year of data. The total number of observer days that can be afforded is 4,062 which represents a 33% increase from 2017. Gap analyses that examine the chance of at least one or three observed trips in a NMFS Area × gear type combination (cell) were used as a performance metric. Gap analyses illustrated that stratifications based on gear type (3 strata) were outperformed by stratifications based on gear type × tendering activity (6 strata). Potential gaps in observer coverage appear to be mostly concentrated in areas with low fishing effort with fewer than 12 trips in a cell. Simulations were performed to measure the potential impact of unknown vessel participation in electronic monitoring (EM). The variability in gap analyses from randomized differences in EM participant vessels was relatively minor (less than 10% probability of observation shifts across deployment designs). The NMFS recommended an observer deployment design for the draft 2018 ADP that has gear type × tendering stratification and uses a "hurdle" approach to sample allocation wherein 15% base coverage is obtained first across all strata and the remainder is optimized according to the variance in the metric of discarded groundfish catch combined with PSC Pacific halibut and Chinook salmon. At their October 2017 meeting the Council did not support the NMFS recommendation and instead proposed a five strata design with optimal sample allocations based on discarded groundfish catch and PSC of Pacific halibut only. Comparisons between the NMFS and Council recommended designs were included in the final 2018 ADP. [doi:10.7289/V5/TM-AFSC-364 (https://doi.org/10.7289/V5/TM-AFSC-364)]

Animal Empires: The Perfection of Nature Between Europe and the Americas, 1492-1615 demonstrates how Renaissance patrons, naturalists, and husbandmen developed useful but dangerous ideas to make sense of natural diversity -- nobility, race, and species -- during the consolidation of the sixteenth-century Spanish Empire. Using the three major techniques at their disposal-relocation, cultivation and training, and selective breeding-elites began a colossal experiment. They sought to create an improved version of Christian nature both in European courts and then, on a larger scale than ever before imagined, in the Americas. Case studies focus on breeding theories and practices in Mantua, Naples, Madrid, Peru, and the Valley of Mexico. Starting in the heart of Renaissance Italy and ending high in the Andes, this project integrates disparate fields of investigation -- ranging from Renaissance aesthetics and animal studies to the histories of the Spanish Empire and of biology -- to reveal an ideal of nature grandly envisioned and prosaically enacted through imperial conquest.

Mechanistic insight into nuclear reprogramming from one cell state to another is of fundamental and clinical importance. Here we use the heterokaryon cell-fusion model of nuclear reprogramming to capture the dynamic architecture of chromatin accessibility and gene expression. At the onset of reprogramming, we detect a transient, genome-wide increase in accessible sites enriched for the AP-1 transcription factor motif assayed by ATAC-seq. Inhibition of AP-1 by expression of a dominant-negative results in an increase in OCT4 expression in heterokaryons. Moreover, in human iPSC reprogramming, dominant negative AP-1 can replace exogenous OCT4 in the reprogramming cocktail. We identify c-Jun as the inhibitory AP-1 family member and show that its repressive activity is mediated through interaction with NURD complex component MBD3 in a phosphorylation dependent manner. Our findings reveal that AP-1, which is induced at the onset of reprogramming and traditionally thought to be an activator, creates a JUN-MBD3 repressor complex that inhibits nuclear reprogramming to pluripotency through direct targeting of an OCT4 distal regulatory element. These findings reveal an unexpected role for Jun as a repressive epigenetic gatekeeper of reprogramming to pluripotency.

This thesis covers several applications of Operations Research in the domains of finance and healthcare. There are three chapters, each covering a different application. Chapter 1 applies techniques from deep learning to estimate mortgage risk. The near-elimination of feature engineering is substituted by models with several thousand parameters that require large amounts of training data. The predictive performance of these models strongly exceeds that of baseline models, especially for predicting prepayments. This increased accuracy, however, comes at the cost of a more opaque model which is harder for a human to interpret than simpler models like logistic regression, which are currently the industry standard. The work in Chapter 2 explores the design of policies for biometric authentication. It first develops a model for the joint distribution of similarity scores associated with different fingers and irises. In the second step, this model is harnessed to design near-optimal multi-stage policies that would be used for authentication, and are robust to gaming, can be computed in real-time and are personalized for optimal performance. The work shows that a reduction of several orders of magnitude in the error rates is achievable by solely changing the authentication policies -- and leaving the hardware unchanged. Chapter 3 is motivated by a humanitarian cause geared towards helping developing countries -- to reduce mortality in children by identifying effective interventions at the planning stage. It takes a descriptive health model (called LiST), which estimates mortality of children given coverage of interventions, and embeds that into an optimization engine in order to minimize mortality under a fixed budget. In doing so, it allows LiST to be used in a prescriptive framework, where policymakers can identify the optimal intervention set at a fixed budget as well as recognize the trade-off of mortality reduction and budget allocation. We find that a greedy strategy offers near-optimal performance with ease of implementation. The findings also highlight the critical role that optimization plays in mortality reduction.

Due to a convergence of large open-source datasets, significant improvements in the parallelization capabilities of hardware (notably, multi-thousand core graphics processing units), and renewed academic interest in decades-old neural network algorithms, primary subfields of AI have flourished in the past 5 years. Natural Language Processing, Computer Vision, and robotics have attained impressive performance across many key AI tasks. In this thesis we focus not on the development of new AI algorithms, but on their application to a series of important problems in healthcare and medicine. Machine learning and AI will impact industries that generate significant amounts of data, by extracting insights that humans cannot. Healthcare and medicine are examples of such industries and thus are excellent use cases for AI deployment. Our story is divided into 4 sections - neuroscience, psychiatry, drug screening, and dermatology - all linked by the common thread of using AI to \textit{enhance the expert}, either in-clinic or in the analysis of data. This underlying motif is the connection to a paradigm in AI development popularized as the \textit{virtual cycle of AI}: build a product that has front-facing user value, generate data with it, use that data to train AI algorithms that improve your product, acquire more data, etc.

Fully automated driving will require intelligent systems capable of understanding, reacting to, and interacting with the intricate complexities of the real world. With the onset of autonomous driving it becomes increasingly necessary to develop advanced tools for establishing trust in intelligent safety systems that act without or despite human input. This thesis presents novel contributions to simulation-based validation, including human driver behavior and sensor models, distributions over driving scenes, and a new technique for the accelerated validation of advanced automotive active safety systems such as autonomous vehicles. Advances to human driver behavior models include the introduction of behavioral cloning models based on Bayesian networks that better capture driver behavior over short horizons. A general input architecture for deep sensor models is introduced and used to develop a stochastic model over an automotive radar's power field. Original contributions are made to the representation of distributions over driving scenes and situations, which must capture a variable number of traffic participants on arbitrary roadways. Finally, this thesis introduces a new method for accelerated validation using importance sampling over clusters of critical situations, prioritizing simulation of critical scenes and avoiding countless simulations of benign driving scenarios while backing out the correct performance statistics.

In this thesis I use physically accurate, end-to-end camera simulation environment to explore different imaging system architectures and to co-optimize the acquisition hardware with computational imaging and computer vision algorithms. The simulation tools use 3D modeling to create different scene models. Scene images that are projected onto the sensor are computed with ray tracing tools. Finally, realistic sensor and camera pipeline simulators compute pixel values that can be used by subsequent computational photography or vision algorithms. I use the simulation tools to optimize the spectral characteristics of cameras and lights used to collect data for inverse estimation methods that characterize surface spectral properties. I show that appropriately selected narrowband lights improve the performance of pixel based surface classification. I also propose a new algorithm that extends beyond narrowband light selection by computing the spectral power distribution of light that is optimal for such classification tasks. I further use simulations to guide and optimize the design of a system that characterizes fluorescent materials which interact with incident light in more complex ways than simple reflection. Spectral illuminant optimization is also useful in consumer photography. I describe a spectrally tunable flash and computational algorithms that can be used to efficiently estimate the ambient illuminant spectrum and to improve the color reproduction of captured images. Such flash is particularly useful when capturing images under extreme ambient lights, for example in underwater photography. The end-to-end nature of the simulation tools makes it possible to explore how viewing conditions (e.g. light level, spectral content, depth), hardware components of an imaging system (e.g. optics, sensor QE, filters, pixel and noise properties) and image processing pipelines affect both the perceived image quality of camera images and the ability of machine learning algorithms to detect and classify objects. I use simulation to render large collections of images in order to evaluate the object detection performance of convolutional neural networks (CNNs) . I demonstrate that the performance of CNNs trained using images that are generated by simplified image processing pipelines (ISP) is similar to the performance of CNNs trained using images generated by more complex and time-consuming ISPs optimized for consumer photography. I also explore and quantify the robustness and performance bounds of detection methods against fundamental algorithms controlling camera exposure and focus.

Enantioselective dihalogenation is an important method for the total synthesis of stereocomplex polyhalogenated natural products, but only a few methods with significant limitations currently exist for this transformation. In 2015, we developed the chemo-, regio- and enantioselective bromochlorination of allylic alcohols featuring a titanium half-salen catalyst to help address this problem. The utility of this method has been demonstrated by the enantioselective total syntheses of 11 halogenated natural products to date by our lab, each of which was directly enabled by this method. Following development, we applied the enantioselective bromochlorination to various new substrates including homoallylic alcohols, which resulted in the total synthesis of (--)-anverene, a polyhalogenated marine natural product with modest but selective antibiotic activity. Computational and experimental mechanistic work was undertaken to better understand the catalytic enantioselective bromochlorination, as well as the enantiospecific solvolysis of resulting enantioenriched bromochlorides. The enantioselective dihalogenation catalyst system has also been extended to new reactions using other electrophile-nucleophile pairs, including haloazidation, which uses TMSN3 as the nucleophilic azide source.

The anthropogenic and the natural biosphere components of the carbon cycle play critically important roles in determining the future status of earth's climate. While our knowledge of CO2 fluxes, at large (global) and small (~1 km2) scales are fairly well known, much uncertainty exists at intermediate/regional-scales (i.e., sub-continental, biome-level). Knowledge at these regional scales is vital for informing relevant climate change mitigation strategies (e.g., land management, fossil fuel emissions reduction policies) and for improving the characterization of key carbon-climate feedback mechanisms. This dissertation leverages the information content of a recent 4-fold increase in atmospheric CO2 observation locations over North America to investigate carbon fluxes, both anthropogenic and natural, at the regional scale (e.g., biome/sub-continental). This work specifically emphasizes the exploration and characterization of spatial and temporal patterns of surface fluxes rather than focusing solely on carbon budgets. With the growing interest in independently verifying fossil fuel CO2 (FFCO2) emissions using atmospheric CO2 emissions, there exists a crucial need to understand the capabilities of the atmospheric observation network to isolate the FF signal. The first study explores and identifies when and where the space-time patterns of monthly sub-continental FFCO2 emissions are detectable using the atmospheric CO2 observation network. Winter months and regions with a relatively high density of observations (e.g., Midwest) offer the best opportunity to detect FFCO2 emissions patterns. The combined impact of the natural biospheric signal and atmospheric transport model related issues are identified as severely hampering the detection of the FFCO2 signal in spring, summer, and autumn. This first study provides key guidance for future efforts to independently estimate FFCO2 emissions using atmospheric CO2 observations through a systematic examination of the various factors hampering detectability. The second study quantifies the ability of a promising new remote sensing data set, solar-induced fluorescence (SIF), to inform spatiotemporal patterns at previously unexplored regional scales. The potential of SIF lies in its ability to measure an emission that, unlike previous reflectance based vegetation indices, is directly related to photosynthesis rather than a measure of the "greenness" of the land surface. While SIF has been explored at local and global scales, here, SIF is shown to explain CO2 flux patterns within continental and biome regions better than existing vegetation and climate indicators as well as most process-based terrestrial biosphere models (TBMs). By incorporating SIF into an inverse model, SIF is found to inform a significant increase in the net uptake of CO2 over croplands as well as a significant decrease over needleleaf forests. The final study explores the regional drivers of interannual variability (IAV) for the net North American CO2 land sink using inverse estimates of net CO2 flux derived from six years of observations. Understanding the current drivers of IAV is crucial for improving predictions of the future response of the terrestrial biosphere to climate change. This work identifies the deciduous broadleaf and mixed forest biomes as the primary regional drivers of IAV over North America, which differs from the dominant drivers identified for the globe and for the northern hemisphere. When comparing the inverse-modelling-derived estimates to a suite of ten TBMs, the large spread in TBM based biome-level contributions to North American IAV make identifying a dominant biome-level driver not possible. The wide spread among TBM based biome-level IAV contributions is attributed to a trade-off between the contribution of IAV in forested vs non-forested regions in a given TBM. This trade-off corresponds with emergent regional sensitivities to environmental drivers (temperature, precipitation, and radiation) where TBMs with IAV dominated by forested regions exhibit stronger sensitivity to environmental drivers in forested regions relative to non-forested regions and vice-versa. This trade-off helps to explain the inability of TBMs to agree on a dominant regional driver of IAV and calls into question the ability of TBMs to inform regional-scale carbon flux IAV dynamics.

Transglutaminase 2 (TG2) is a ubiquitously expressed, Ca2+-activated enzyme in mammals that catalyzes the formation of cross-links between glutamine and lysine residues on protein or peptide substrates. While chiefly a cytosolic enzyme, TG2 can localize to the extracellular environment through a poorly understood non-classical secretory mechanism. Aberrant extracellular TG2 activity has been implicated in several human diseases. Most notable is celiac disease (CD; celiac sprue), a widespread lifelong autoimmune disorder that affects the small intestine and is driven by the consumption of dietary gluten. In the context of CD, TG2 is responsible for the deamidation of immunogenic gluten-derived peptides, resulting in their increased antigenicity in genetically susceptible individuals (HLA-DQ2 or HLA-DQ8). However, extracellular TG2 is predominantly catalytically inactive in most organs under normal physiological conditions and the precise mechanism of activation in the celiac gut is unknown. A clearer understanding of how TG2 activity is regulated at the post-translational level could shed light not only on our understanding of celiac disease pathogenesis but lead to the development of novel therapeutic strategies to combat diseases affected by abnormal extracellular TG2 activity. In the first part of this thesis, we identify a novel class of compounds that exhibits unusual dual antagonist and agonist action on TG2. Acylideneoxoindoles were first identified in our lab as reversible inhibitors but closer examination has revealed that these molecules behave as activators under sub-saturating but physiologically relevant Ca2+ concentrations. Detailed analysis of a lead compound, CK-IV-55, revealed that these class of molecules target low-affinity Ca2+ binding sites on the catalytic core of TG2. This discovery sheds light on potential dietary triggers of TG2 activation as indoles are abundant in cruciferous vegetables and can also be produced by commensal gut bacteria. Until recently, the mechanisms that lead to inactive extracellular TG2 remained a fundamental mystery; the extracellular environment fosters conditions that favor constitutive TG2 activation due to the abundance of Ca2+. Identification of a vicinal disulfide bond (Cys370-Cys371) that inhibits enzymatic function and acts as a protein redox switch has provided the missing link. While disulfide bonds play a large role in maintaining tertiary and quaternary protein structure, allosteric disulfide bonds that control the function of mature proteins have recently been identified. Previous efforts in our lab have established that the redox protein cofactor thioredoxin-1 (TRX) could switch 'on' TG2 in vitro and in vivo through cleavage of the vicinal disulfide bond but it is unclear how TG2 is switched 'off.' Here, we systematically evaluated biologically relevant oxidants for their ability to oxidatively inactivate TG2 and identified the thiol-disulfide oxidoreductase, ERp57, as a suitable candidate. This discovery presents the first example of an allosteric disulfide bond redox switch that is dynamically regulated by two distinct proteins. Lastly, we validate the redox regulation of human TG2 through mutagenetic analysis and develop a robust tissue culture model that displays constitutive extracellular TG2 activity. This powerful tool overcomes limitations of previous models and could elucidate the biological consequences of extracellular TG2 activity. In summary, this dissertation provides insights into the complex post-translational regulation of TG2 using a variety of chemical biology techniques. We identify acylideneoxoindoles as allosteric chemical activators of TG2 that can be derived from the diet or other exogenous sources. We also provide a mechanism for the oxidative inactivation of TG2 through a unique redox switch controlled by two distinct proteins. Lastly, tools for developing robust cell culture models to assess the biological consequences of extracellular TG2 activity are discussed. These findings serve as the much-needed foundation to understanding the pathophysiological implications of this enigmatic protein.

"This NOAA Technical Memorandum represents a 2017 update to the Commonwealth of the Northern Mariana Islands as a Fishing Community published by Allen and Amesbury in 2012. This fishing community profile update aspires to provide a comprehensive view of the Commonwealth of the Northern Mariana Islands (CNMI) as a fishing community. The report begins by presenting recent social, economic, and tourism information gathered by the Government Accountability Office (GAO) and the Marianas Visitors Authority in the CNMI. These data are followed by data on four key topics that emerged from interviews conducted by the author in 2016 with the CNMI fishing community. Next, findings from low income fish consumption in 2015 are presented, aggregated at the island level. These data provide the only current estimates of fish consumption on Tinian and Rota. The report concludes by presenting some additional empirical information that supplements the primary data gathered in the profile"--Executive Summary. [doi:10.7289/V5/TM-PIFSC-66 (https://doi.org/10.7289/V5/TM-PIFSC-66)]

Secondary lithium (Li) ion batteries are essential in driving the rapid development of electronic devices. While the rising demand for high performance portable electronics continues to sustain interest in developing more advanced lithium ion (Li-ion) batteries, the emerging applications of grid scale energy storage and electric vehicles are pushing lithium battery research to the next level. To meet the targets for grid storage as well as electric vehicles batteries set by the US Department of Energy (DOE), battery chemistries beyond the current lithium ion systems are highly required. Among all the recently-emerging technologies, lithium sulfur (Li-S) battery is one of the leading candidates that could have high specific energy and low cost. In my thesis, I will very briefly introduce the fundamentals of conventional Li-ion batteries and the next-generation Li-S batteries, compare the different chemistries of conventional lithium-ion batteries with Li-S batteries, and then examine the main challenges and present my study on designing the nanoscale "composite electrode" to address problems from both the lithium metal side and sulfur side. Lithium sulfur battery has a practical energy density of around 600 Wh/kg, about 3 times that of the current lithium ion batteries. Lithium metal anode has long been regarded as the "holy grail" of battery technologies, due to its low potential and high specific capacity. However, the safety hazards and capacity fading problem have prevented lithium metal anode from practical realization. In the first part of this thesis, I will present my research on using various strategies to build a composite Li metal anode. By encapsulating Li inside a three-dimensional porous matrix via melt-infusion, electrochemical deposition and mechanical deformation, we demonstrated the fabrication of such a composite anode. The resulting lithium--matrix composite anode was then subjected to battery cycling and exhibited superior performance compared with bare lithium metal anodes. In the last part of this thesis, I will describe my work on using hydrogen reduced titanium dioxide nanostructures as matrix to improve the sulfur cathode performance. On the cathode side, sulfur and its discharge product are highly insulating. In addition, the intermediate discharge products (lithium polysulfides) can easily dissolve into the electrolyte and shuttle between electrodes, which leads to a fast capacity decay and limited cycle performance. By embedding sulfur in hydrogen reduced titania inverse opal structure, sulfur utilization in the electrode is significantly improved and the polysulfide species are strongly trapped both physically and chemically, resulting in higher specific capacity and longer cycle life. We believe our work will contribute significantly to the energy-related field and also inspire research in other areas.