UCLA

A general method incorporating both vibrational and configurational degrees of freedom is proposed by expanding energy into vibrational clusters. Our approach captures vibrational properties in substitutional alloys with arbitrary atomic configurations which can serve as an accurate surrogate model for first-principles calculations. Compressive sensing is applied to robustly and accurately select the important configuration dependent force constants and determine their value by learning from first principles calculations in one shot. Unlike virtual crystal approximation which tends to overestimate lattice thermal conductivity at high concentration range. Great agreements with experimental results across all composition for PbTe-Se demonstrates our VCE method is an accurate approach to generate high fidelity potential energy surface across a wide range of alloy materials.

This dissertation contains three essays in Applied Microeconomics. Chapter 1 provides the first causal estimates of the effect of children’s access to computers and the internet on adult educational outcomes such as schooling and choice of major. I exploit cross-cohort variation in access to technology among primary and middle school students in Uruguay, the first country to implement a nationwide one-laptop-per-child program. Despite a notable increase in computer access, educational attainment has not increased. However, college students who had been exposed to the program as children, were more likely to select majors with good employment prospects. Chapter 2 provides the first empirical evidence of the historical effects of natural disasters on economic activity in the United States. Although the literature has focused on salient natural disasters, more than one hounded strike the country every year, causing extensive property destruction and loss of life. My coauthors and I construct an 80 year panel data set that includes the universe of natural disasters in the United States from 1930 to 2010 and study how these shocks affected migration rates, home prices and poverty rates at the county level. Severe disasters increased out-migration rates by 1.5 percentage points and lowered housing prices/rents by 2.5–5.0 percent, but milder disasters had little effect on economic outcomes. Chapter 3 exploits the 1962 publication of Silent Spring, the first successful environmental science book, to investigate whether public information can influence popular demand for environmental regulation. Protecting the environment is often plagued by collective-action problems, so it is important to understand what motivates politicians to act. Combining historical U.S. congressional roll-call votes and census data, I find that the propensity of politicians to vote in favor of pro-environmental regulation increased by 5 to 33 percentage points after the publication of the book. The response to the informational shock varies with the constituency’s level of education, income, and exposure to pollution.

Choreographing Livability: Dance Epistemes in the Kibbutz and in the Israel Defense Forces traces the historical articulation of dance as a source of knowledge-formation in Israeli culture through two emblematic sites of performance, between the 1940s and the 2000s. It also proposes a theoretical intervention through the elaboration of the framework of livability, through which I explore the life-stakes and the political investment entailed in dancing within the specific context of Israel, in relation to its larger ideological tensions and political shifts.

My investigation across sites of performance and time-periods ultimately reassesses existing narratives that have framed “Israeli dance” primarily as a joyful, nation-building, recreational, entertaining, and energetic endeavor. In order to do so, I set out the mechanisms through which different dance experiences, even those apparently disengaged from political preoccupations, have contributed to the enhancement of governmental policies and ideological goals, in particular when such political maneuvers reiterated ethnonational divides or mechanisms of settler colonial hegemony. More specifically, through my scrutiny, supported by archival research, ethnography, and choreographic analysis, I unpack how dancers and choreographers in Israel have often articulated dance as a multicultural, universalistic, and humanizing practice. By doing so, I maintain, dance in Israel has generally worked as a strategy for the mitigation and concealment of larger governmental and ideological apparatuses of marginalization, commodification, or oppression.

The Introduction offers an interpretation of Zionism and Israel from a biopolitical perspective, an overview of my livability framework, and a reading of my project in terms of killjoy scholarship. Chapter 1 delineates how dance in kibbutz culture has been able to support shifts in the national strategy, evolving from engine for the international affirmation of Zionism, to agent for a rearticulation of the Socialist Labor Zionist agenda, to neoliberal enterprise. Chapter 2 charts the evolution of dance in the Israel Defense Forces from bureaucratic tool for the administration of military life, to spectacular device for the recalibration of the Israeli soldier’s masculinity, to globalized digital practice that reinforces military authority from the lower levels of the military hierarchy. The Epilogue, in addition, includes four choreographic analyses that, engaging with the kibbutz, the IDF, and the issue of choreographing in Israel, show how dance can invest in a critique of systems of oppression, and expand the possibility of living more livable lives.

This dissertation addresses various engineering problems involving the seismic re- sponse modeling of earth-retaining structures. These are namely, (i) lateral pas- sive seismic behavior of ordinary skew-angled bridge abutments, (ii) lateral pas- sive seismic behavior of high-speed rail transition abutments (with no skew), and finally (iii) active and passive seismic behavior of (cantilevered) earth-retaining walls. The approach adopted for each problems is the same, which is to devise a macroelement model with physics-based parameters (e.g., soil density, shear strength, wall height, etc.) that captures salient response features. These models are able to predict the lateral capacity of the retained soil and residual displace- ments with a modest computational effort—as compared to, for example, predic- tive simulations carried out with three-dimensional finite element models—, which renders them to be amenable for repeated nonlinear time-history analyses required for performance-based seismic assessment and design. The three aforementioned problems are briefly described below:

I. Presence of skew-angled abutments complicates the seismic behavior of or- dinary bridges, primary driver of which is the passive lateral resistance of the engineered backfill behind the abutment. The eccentricity of the soil reaction

ii

relative to the bridge’s center of stiffness or mass causes a skew bridge to rotate under seismic excitations, and a nonuniform soil pressure distribution develops behind the abutment backwall. A distributed nonlinear spring model is devised here to represent the lateral passive reaction of the backfill soil. To that end, a modification factor is devised so that Log-Spiral Hyperbolic (LSH) backbone curves –which had been developed in prior research and were validated for back- fills of straight abutment–can be used to generate the backbone curves of the said springs. This new modeling approach is verified against three-dimensional finite element model simulations and is validated with data from large-scale experiments conducted at Brigham Young University that had produced direct measurements of load-deformation backbone curves for several skew angles. In the final step, the validated modified-LSH model is used in parametric studies to devise a simple bilinear load-deformation relationship that is parameterized with respect to the backwall height, abutment skew angle, and the backfill soil properties. This sim- ple relationship is intended for routine use in the capacity-based seismic design and analysis of skew bridges.

II. California’s High-Speed Rail (HSR) System is slated to traverse nearly the entire length of the state, and thus it will be exposed seismic risks from almost every known major tectonic fault there. The present study deals with the seismic responses of bridge-abutment transition backfills (BATBs), which are essential components of HSR bridges. BATBs provide a gradual variation of vertical stiff- ness between the bridge deck and the engineered backfill zone, enabling smooth operations for trains traveling at high speeds. All prior investigations focused on this vertical stiffness in order to better characterize the localized vertical dif- ferential movements around BATBs under periodic high axial loads from train sets. Lateral behavior of BATBs, which are important under seismic loads, have not been previously investigated. The present study offers a parametric nonlin- ear lateral force-displacement backbone curve for BATBs that is verified against

iii

three-dimensional finite element models and validated against data from large- scale tests conducted at Brigham Young University. The parametric curve takes backwall height as well as abutment skew angle into account.

III. Performance-based seismic assessment (PBSA) of earth retaining struc- tures requires the use of accurate yet computationally efficient analysis models. To date, limit equilibrium models offered the most computationally efficient re- sults, but they only produce estimates of peak lateral seismic forces and cannot be used in nonlinear time-history analyses. While detailed finite element models can possibly fill this need, they are not amenable for repeated simulations required for quantifying the uncertainties associated with estimated ground motions within the PBSA framework. A novel Lumped Impedance Model (LIM) is developed in this study that generates as accurate solutions as detailed FE models, with trivial computational effort. The model is able to also reproduce lateral passive load-deformation backbone curves as predicted by a state-of-the art limit equi- librium model, by its design. The computational saving offered by LIM is due to lumping of mass and stiffness of the retained soil, and the strategic placement of elastoplastic macroelements along pre-calculated active and passive failure hy- perplanes. LIM is verified against analytical solutions in frequency-domain for linear response regimes—wherein it is shown that LIM can accurately capture the frequency-dependent responses of the retained soil—as well as other previous studies for inelastic conditions. LIM is also verified against detailed FEM sim- ulations of cantilevered retaining wall subjected to both narrow- and broadband excitations, and it is shown that both elastic and inelastic responses of the retained soil (including residual wall displacements and rotations) are adequately captured. Finally, a framework for PBSA of earth-retaining structures using LIM as the pre- dictive model is proposed and its use is demonstrated through an example seismic assessment application wherein a fragility curve is computed.

iv

The present study replicates a 1D, steady, chemically-reacting, reduced-order model for hydrogen peroxide flow through a monopropellant catalyst bed as described in Pasini et al. [9] with model validation completed by comparison with both model data and experimental data from Jung et al. [10]. Adaptations were made to improve heat transfer capabilities within the model and to adapt the model such that a hydroxylammonium nitrate and water mixture could be used as the propellant. Polynomial chaos expansion was implemented to decrease sampling time in order to perform non-deterministic analyses including: quantification of global sensitivities using Sobol indices, construction of axial property profiles with uncertainty envelopes for random physical inputs, and construction of posterior probability distributions with confidence intervals for variation in chemical tuning parameters. Results from the study show that model behavior is primarily governed by propellant mass fraction and activation energy of the global reaction. Additionally, posterior distributions indicate that activation energy and number of active sites per volume are related by a logarithmic family of solutions as a result of the reaction advancement gradient form in the model.

Graph structures are a powerful abstraction of many real-world data, such as human interactions and information networks. Despite the powerful abstraction, graphs are challenging to model due to the high-dimensional, irregular and heterogeneous characteristics of many real-world graph data. An essential problem arose is how to effectively and efficiently learn the representation for objects in graphs. In this thesis, both the effectiveness as well as efficiency aspects of the graph representation learning problem are addressed. Specifically, we start by proposing an effective approach for learning heterogeneous graph embedding in an unsupervised setting. Then this is generalized to semi-supervised scenario where label guidance is leveraged. The effective graph representation learning models are followed by efficient techniques, where we propose efficient sampling strategies to improve the training efficiency for content-rich graph embedding models. Finally, to reduce storage and memory cost of the embedding table used in various models, we introduce a framework based on KD code, which can compress the embedding table in an end-to-end fashion. We conduct extensive experiments on various real-world tasks on graph data (e.g. anomaly detection, recommendation and text classifications), and the empirical results validate both effectiveness as well as efficiency of our proposed algorithms.

One of the holy grails in cancer therapy is to simultaneously image and deliver drugs to the tumor site. The first part of the thesis has developed new ideas in cancer theranostics and the second part is about the development of a novel contrast agent for risk-free imaging of the tumor. In the first project of the thesis, I have discussed the development of a liposome-based cargo delivery strategy that can simultaneously monitor alternating magnetic field–induced drug release by observing the change in MRI relaxation parameter R1, and the location and condition of liposomal site (such as tumor) from MRI parameter R2. However, the loading of a contrast agent in liposomes generally results in poor contrast and suffers from various artifacts in in vivo experiments, compared to the use of free contrast agent. Thus, the second project of my work in this part demonstrates the effective filtering of artifacts and contrast enhancement to obtain high quality sensitive imaging of the tumor site in a mouse model using paramagnetic liposomes, a novel pulsing sequence in active-feedback MRI, and robust data analysis. The second part of my work is about using a self-replicating viral-RNA molecule derived from Nodamura, an insect virus, to express and amplify ferritin, leading to increased iron content of the cells in the form of ferrihydrite that acts as a novel contrast agent. In summary, my thesis is about the development of cancer theranostics and a novel contrast agent. In the last part of the thesis, I have reviewed the current development of rapidly growing state of the art magnetic resonance cancer theranostics for commonly used polymer-based nanovehicles.

For many types of animal cells, migration is essential for their physiological function. Long distance crawling is required for the development of a nervous system and for the ability of macrophages to accumulate at sites of infection. In response to a wound, resident fibroblasts migrate to the wound site and promote the wound healing process. Furthermore, cancer metastasis involves the ability of cancer cells to transform their morphology and acquire an ability to migrate. A better understanding of cell migration would have implications for normal development and could yield therapies for multiple different disease states, including fibrosis, atherosclerosis, and metastasis.

Alternative polyadenylation (APA) is the process by which the same genetic locus can produce multiple transcripts ending at different polyadenylation sites. Coding sequence APA can change the length of the amino acid and can also switch the function of the protein. Proliferating cells, cancer cells, and stem cells in the early stages of mouse development and tumors tend to use proximal polyadenylation sites and commonly show rapid cell migration. Our laboratory has discovered that fibroblast migration can be regulated through polyadenylation site selection. To date, however, no studies have been performed to research how cell migration and invasion can be controlled by APA.

Reversion-inducing-cysteine-rich protein with kazal-like motifs (RECK) suppresses cell migration. Previous studies have documented a role for RECK on the cell surface in inhibiting MMP activity via its Kazal-like motifs. RECK has also been implicated in tubulin dynamics and in anterior-posterior polarity. Our lab discovered through an RNA-Seq-based screen that there are short and long RECK isoforms that are generated by coding region alternative polyadenylation.

In this dissertation, I explain how cell migration and invasion can be controlled by APA via alternative isoform expression of the RECK protein. I show that the short RECK isoform promotes cell migration and that this result is mediated by short RECK and long (canonical) RECK interaction. Furthermore, I propose that RECK APA can regulate cell migration not only through its established role of modifying matrix metalloproteinases but also via its roles in affecting tubulin post-translational modifications (PTMs). These studies provide a much more detailed and mechanistic model for the biological functions of RECK alternative isoform use.

Although it is often assumed that production must logically precede consumption, the development of postmodern architecture complicates this narrative. The development of postmodern architecture undermined established structures by centralizing the role of consumption and the consumer. This dissertation examines ways in which various conservative trends pushed the consumer closer toward production. These changes ushered in the experience economy of which Best Products Company and the broader catalog showroom phenomena were particularly emblematic. Drawing on these changes within the history of retail architecture, this dissertation sets out to explore how architecture emerged into the postmodern period as a box, a malleable shell that was increasingly being invaded and overturned by a powerful consumer.

Ubiquitous smart devices and internet of things create tremendous data every day, shifting computing diagram towards data-driven. Computing and memory units in traditional computers are physically separated, which leads to huge energy cost and time delay. Novel computer architectures bring computing and memory units together for data-intensive applications. These memory units need to be fast, energy efficient, scalable and nonvolatile. This dissertation concerns innovating new types of magnetic memory or spintronic devices to achieve ultrahigh energy efficiency and ultracompact size from a perspective of material and heterostructure design. Especially, we employ quantum materials to enable potentially unprecedented technological advances. The highest energy efficiency of magnetic memory requires the largest charge-to-spin conversion efficiency that allows the minimum power to manipulate the magnetization. We utilize topological surface states of topological insulators (TIs), which have unique spin-momentum locking and thus are highly spin-polarized. We discover giant spin-orbit torques (SOTs) from TIs at room temperature, which are more than one order of magnitude larger than those of traditional heavy metals. We integrate TIs into room temperature magnetic memories, which promises future ultralow power dissipation. SOT characterization methods and related SOT studies on heavy metals, monolayer two-dimensional materials, and magnetic insulators-based heterostructures are discussed in detail. To have the best scaling performance, we investigate emerging topological skyrmions in magnetic thin films, which are arguably the smallest spin texture in nature. While most of the skyrmions are discovered in metallic systems, insulating skyrmions are desired thanks to their lower damping and thus potentially lower power dissipation. We observe high-temperature electronic signatures of skyrmions in magnetic insulators, topological Hall effect, by engineering heterostructures consisting of heavy metals and magnetic insulators. This new platform is essential for exploring fundamental magnon-skyrmion physics and pursuing practical applications based on insulating skyrmions. To have the highest operation speed, we explore compensated ferrimagnetic insulators, which have THz dynamics due to the strong exchange coupling field. We realize energy efficient switching of the ferrimagnetic insulator in both ferrimagnetic and antiferromagnetic states, promising electrical manipulation of ultrafast dynamics.