Contributors: Wang, Peiqi, Carmona, René, Operations Research and Financial Engineering Department
... Mean field game is a powerful framework for studying the strategic interactions within a large population of rational agents. Although existing research has predominantly relied on diffusion models to depict agents’ states, numerous applications, such as epidemic control and botnet defense, can best be modeled by systems of particles in discrete state space. This thesis tackles finite state mean field games. In the first part of the thesis, we develop a probabilistic approach for finite state mean field games. Based on the weak formulation of optimal control, the approach accommodates the interactions through the players’ strategies and flexible information structures. The second part of the thesis is devoted to finite state mean field games involving a player possessing dominating influence. Two different mechanisms are explored. We first study a form of Stackelberg games, in which the dominating player, referred to as principal, moves first and chooses its strategy which impacts the dynamics and ob- jective functions of every remaining player, referred to as agent. Having observed the principal’s strategy, the agents reach a Nash equilibrium. We seek optimal strategies of the principal, whose objective function depends on the statistical distribution of the agents’ states in equilibrium. Using the weak formulation of finite state mean field games developed previously in the thesis, we transform the principal’s optimization problem into a McKean-Vlasov control problem, and provide a semi-explicit solution under the assumptions of linear transition rate, quadratic cost and risk-neutral utility. In the second model, we assume that all players move simultaneously and we study Nash equilibria formed jointly by major and minor players. We introduce finite player games and derive mean field game formulation in the limit of infinitely many minor players. In this limit, we characterize the best responses of major and minor players via viscosity solutions of HJB equations, and we prove existence of Nash equilibria under reasonable assumptions. We also derive approximate Nash equilibria for the finite player game from the solution of the mean field game.
Contributors: Mack, Jessica Robin, Adelman, Jeremy, History Department
... In June of 1950, the first stone was ceremoniously placed on a construction site several kilometers south of Mexico City’s center. Building had officially begun on a long-imagined campus for the National Autonomous University of Mexico (UNAM): the utopian Ciudad Universitaria, or University City. By tracing its spatial reconfiguration on a new modernist campus, this dissertation explains the national university’s shifting role in Mexico’s developing revolutionary state and examines the ways in which new national priorities were inscribed upon intellectual and cultural life. The largest university in Latin America, with over 330,000 students today and a sprawling 2,500-acre urban campus, UNAM has trained Mexico’s most powerful decision makers, set the terms of public debates and posed questions that have shaped decades of policy. It is also an important channel through which citizens have participated—with both consent and contestation—in Mexico’s twentieth-century public life. Although it is an icon of modern Mexico, we know surprisingly little about UNAM’s historical role in shaping the social landscape and the reproduction of power. From the earliest imaginings of a utopian lettered city immediately after the Mexican Revolution (1910-1920), this project traces the long twentieth-century arc of Ciudad Universitaria’s creation alongside the national university’s turbulent relationship with the revolutionary state. As the prominent modernist architects of Ciudad Universitaria solidified the ruling party’s interpretation of the revolution in stone, social justice and agrarian reform were replaced by a new set of priorities. In conversation with a growing body of scholarship on Mexico’s midcentury, the dissertation explores how this monumental state building project spatially manifested the ruling party’s vision for Mexico’s future and was made possible by its new priorities of urbanization, industrialization and developmentalism. From imagining and planning the campus in the late 1940s to its rapid construction from 1950-1954, the dissertation uses the archives of planners, architects, local residents and bureaucrats to reveal new modes of operating. Once students and faculty arrived, the project traces intellectual and social reconfiguration at Ciudad Universitaria, changes that would reshape Mexico City’s urban landscape and realign contestation in national politics into the late twentieth century.
Prediction of Cancer Phenotypes Through Machine Learning Approaches: From Gene Modularity to Deep Neural Networks
Contributors: Zamalloa, Jose Antonio, Singh, Mona, Quantitative Computational Biology Department
... The current genomics data influx is transforming healthcare by enabling precise diagnoses and individualized treatments. This is especially true for cancer, where we have genome sequencing and gene expression data across numerous individuals, along with measurements of drug response across hundreds of cancer cell lines. Computa- tional, statistical and machine learning methods play an essential role in analyzing these data in order to gain medically relevant insights. In this dissertation, I describe statistical and machine learning approaches to enable better stratification of cancer subtypes and predict therapy outcomes for individuals with cancer. First, I introduce Deep Pharmacogenomic Modules (Deep-PGMs), a framework to predict drug response outcomes for tumor samples using drug features and gene expression data. Genome expression signatures are a great aid for predicting whether a particular therapy may be beneficial for a specific cancer tumor. Traditional ma- chine learning approaches to predict the effect of a cancer drug on a tumor typically focus on the expression levels of either certain key cancer-relevant genes or of all genes. While genomic data can aid in describing the disease state of an individual by looking at isolated gene entities, genes in cells tend to act in concert to perform their functions. My approach takes advantage of the modular nature of gene regu- lation to build a reduced feature space that describes the cellular state of a tumor. I take advantage of unsupervised machine learning methods to build genomic and non-genomic feature spaces. I construct a deep neural network pipeline to predict drug efficacy outcomes on tumor cell line samples. I demonstrate that my framework outperforms traditional machine learning approaches that do not take advantage of the modular structure of gene expression data sets. I further apply my method to clinical trial data and demonstrate its performance. I find that featurizing genomic data through prior knowledge about cellulary modularity, accompanied with a robust deep learning pipeline, is a powerful method for predicting the disease outcome of novel cancer therapeutics. In the second part of my thesis, I develop classifiers to identify two breast cancer subtypes. First, I describe an accurate Claudin-low (CL) molecular subtype predictor based on gene expression data. This particular subtype has poor prognosis in breast cancer patients. Via experiments in mice along with analysis of human breast cancer data, my collaborators and I linked individuals with CL breast cancer to elevated lev- els of miR-199a. This evidence further supported the high levels of miR-199a in mice tumors and helped characterize the miR-199a-LCOR-IFN axis in tumor initiation. Next, I developed a hysteretic epithelial-mesenchymal transition (EMT) classifier. I use experimental data from TGF-β induced EMT mouse mammary tumor cells to find genes that are indicative of the hysteretic EMT phenotype. The uncovered genes in my model correlate well with metastatic phenotypes in clinical datasets, particularly in patients with metastatic lung cancer, suggesting that EMT-induced mice mammary tumor cells can help elucidate clinically relevant genes important in metastasis.
Contributors: Sun, Jiming, Car, Roberto, Physics Department
... Water is one of the most common substances in the universe and is found in many different phases. At extremely high temperature and pressure, water ice forms a superionic phase (SI) in which the water molecules dissociate into ions, with the oxygen ions forming a crystal lattice structure and the hydrogen ions flowing through the lattice like a liquid. In this thesis, we present a first-principle study of the high pressure superionic ice phase. We systematically explore the superionic phase diagram and carefully study its physical and chemical properties. As a result of this investigation, we find that superionic ice exists in three different phases differentiating from each other in the structure of the oxygen sublattice. One phase is the BCC-SI phase, which has BCC oxygen sublattice and has been predicted by several groups. We find the BCC-SI phase is stable up to about 250GPa. At higher pressure, BCC-SI transforms via a first order transition to the CP-SI phase. The oxygens form close-packed structures in this phase. CP-SI has also been predicted by other authors, though our predicted transition pressure is significantly higher. Finally, we find a third SI phase, P21/c-SI, which would exist at even higher pressure and it is predicted here for the first time. In addition, we use the imaginary-time path integral technique to study the nuclear quantum effects (NQE). We find the NQE changes the SI ice phase boundaries significantly. Compared to the classical results, NQE makes CP-SI phase more stable and increases the transition pressure from about 1.5TPa to about 1.7TPa. We show that the NQE on the SI phase boundary can be explained by thermodynamic perturbation theory to the lowest order. Lastly, we studied the chemical bonding and the electronic structures of high pressure superionic phases. We computed the distribution of maximally localized Wannier functions centers and showed evidence that the chemical bonding in high pressure ice phases is dominated by ionic interactions. This is very different from the low pressure phases, where the water molecules are intact and the intermolecular interactions are dominated by hydrogen bonds.
Contributors: Santos, Andrew Pablo, Panagiotopoulos, Athanassios Z, Chemical and Biological Engineering Department
... Surfactants and colloids are ubiquitous in biology, the environment, home and personal care products, and industrial applications. The self-assembly of surfactants into micellar aggregates is crucial to their use in many industrial applications, including: detergency, cosmetics, oil-spill remediation and nanoparticle dispersion. Because of their ubiquity, surfactant and colloid self-assembly are canonical subjects of soft matter study, yet there are concentration effects, for example the free surfactant concentration, that remain difficult to measure and understand. We present our findings on the driving forces behind aspects of self-assembly and methodologies for characterizing the behavior of three self-assembling systems: colloids with short-range attraction long-range repulsion, nonionic surfactants and ionic surfactants. To study molecular details of these systems, we use Monte Carlo and molecular dynamics simulations with implicit-solvent models. We directly compare different methods for calculating the critical micelle concentration in all self-assembling systems. We establish that excluded volume and counterion condensation are the main driving forces for the decrease of free surfactant concentration in nonionic and ionic surfactants, respectively. We compare counterion condensation and mean ionic activity measurements directly to experiments, and address experimental disagreements. For short-range attraction long-range repulsion colloids, conditions are found where large preferred aggregates have no net effect on the pressure, which is strikingly different behavior from surfactant self-assembly. The results in this thesis offer a deeper understanding of the phenomena of low-concentration self-assembly, leading to improved methods for estimating micellar properties over wider concentration ranges. These methods will benefit the development of more accurate molecular models and speed up development of new industrial formulations.
Contributors: Ayers, Elaine, Burnett, D. Graham, Milam, Erika L., History of Science Department
... “STRANGE BEAUTY” uses a material culture approach to untangle the botanical dualities that, from the late-eighteenth to the mid-nineteenth centuries, produced a constructed image of East Indian rainforests, obscuring the oftentimes violent racialized, gendered, and sexualized work contained therein. The plants written into scientific and aesthetic descriptions of “tropical nature,” I argue, constructed a view of rainforests as extractable, collectible storehouses caught somewhere between fecundity and loss, between beauty and decay. I focus on four plants transformed from curious objects in the field to herbarium specimens in Britain and Europe bound up in debates over evolution and “normalcy” while speaking to the developing ideals and disillusionments of tropical nature. These objects confused reproductive boundaries, embodying the possibilities and perceived environmental dangers while challenging typical labor practices producing a collectible, useful, and containable nature. Chapter One follows the most miniscule, mundane, of plants: Moss. Paradoxically figured as both “pure” in its seemingly invisible self-reproduction and salacious in its place in pornography, moss functioned as a preservative agent for more “valuable” plants. Simultaneously, botanists interested in moss on its own terms transformed this packing material into an object of microscopic inquiry. Next, I follow an orchid illustrated by a prolific female artist. Explicitly sexualized in their depictions, orchids challenged aesthetic representations, transforming the technical process of botanical illustration while contributing to a false construction of the tropics inevitably leading to affective dissonances among naturalists. Here, theories of color and artistic practice become central to the politics of botanical preservation and display. Chapter Three traces a carnivorous pitcher plant collected by a liminal figure working for the East India Company. The pitcher plant came to crystallize colonial fears of environmental pushback—the phallic, carnivorous plant challenged theories of order while provoking questions about colonial consumption. I end with the largest and rarest plant in the world—the corpse flower. That a flower could actively mimic animal behavior to attract pollinators upended the chain of floral being. Resistant to all forms of collection, preservation, and display, the corpse flower proved to be the ultimate form of tropical nature’s resistance to human intervention and control.
PRIVATE CAPITAL, PUBLIC CREDIT: BRAZIL, ARGENTINA, AND THE PROBLEM OF CREDIBILITY IN INTERNATIONAL CAPITAL MARKETS, 1852—1914
Contributors: Vedoveli Francisco, Paula Elena, Adelman, Jeremy, History Department
... Private Capital, Public Credit examines how Brazilian and Argentine governments promoted their public credit at international capital markets from the 1850s to World War I. It focuses on the relationships that connected sovereign governments to contractors and issuers of foreign debt in London and Paris, and particularly on credibility brokers, a group of actors responsible for establishing and maintaining these connections. Credibility brokers were able to manage the flow of information between decision-makers in borrowing countries and bankers in international capital markets and promoted trusting relationships with actors on both sides of the Atlantic through repeated interactions and the development of affective bonds. Credibility brokers, therefore, shaped financial intermediaries’ capacity to assess sovereign risk and influenced debtors’ ability to make and sustain credible commitments. Their performance was dependent on the actors’ private capital, including the character of the ties that connected them to statesmen in Latin America and bankers in Europe, their ability to adopt the elements that constituted the culture of personal credibility, their capacity to operate independently in capital markets and control the nature of the information available to European investors, and the political capital they enjoyed amongst decision-making elites in borrowing countries. This cosmopolitan model of credibility brokerage started to wane in the last decade of the nineteenth century when financial crises in Brazil and Argentina and the failure of one of the most reputable banking houses in London undermined investors’ trust on the capacity of financial intermediaries to correctly assess risk and monitor sovereign borrowers. The transition to a model of sovereign credibility based on economic indicators and credit-ratings occurred gradually in the first half of the twentieth century. The adoption of this model signaled a growing distrust towards the capacity of personal judgment and knowledge to regulate sovereign bond markets, while the expansion of the investing public and the erosion of actors’ capacity to control the flow of information enabled the change. Credibility brokers, however, have remained a feature of sovereign bond markets, called to perform their roles when the power of numbers has proved insufficient to broker credible commitments in times of heightened uncertainty.
Contributors: Adair, Carl C, Fuss, Diana, Kotin, Joshua, English Department
... Faithful Readings outlines a revisionist history of the discourse and practice of Anglo-American literary criticism between 1880 and 1950. This dissertation contests the secularization narratives that structure many received histories of the discipline, narratives in which habits of critical interpretation and reflection inevitably supersede or simply survive a credulous, pre-modern religion that had become increasingly untenable through the nineteenth century. The project presents a genealogy of this enduring concept of religion as naïve and literal belief, arguing that the concept was produced as a polemical strategy in a theological contest: it was first promoted not by disinterested secular critics but by the under-remarked nineteenth-century theological movement known as Protestant modernism as a dismissive depiction of its theological rivals. For modernists, faith was a self-consciously interpretive relation to the Christian tradition as a complex of historically and culturally mediated forms whose ultimate object was an ineffable, infinite reality: defined against propositional belief, true religion was a way of reading this tradition critically. This dissertation argues that the habits of literary criticism developed in this period were recursively entangled with the powerful and problematic hermeneutics that defined this one contemporary Protestant theology. In chapters dedicated to the reading practices of such formative figures as Matthew Arnold, John Crowe Ransom, Cleanth Brooks, and T.S. Eliot, the project elucidates the political, epistemological, and institutional pressures that led both biblical and literary critics to develop new methods to legitimize their work as specialized, modern disciplines and also to remain faithful to their cherished texts as authorities that exceeded the knowledge such specialized methods could produce. Thinking beyond the purported opposition between the critical and the religious, the project defamiliarizes tensions within enduring habits of literary criticism and invites serious reflection on the forms of discipline that may be required should critics decide to change their habits.
Contributors: Sybblis, Martin, Scheppele, Kim L., Sociology Department
... This dissertation offers a novel way to understand how legal change occurs in the corporate and commercial context. It suggests that legal professionals, particularly commercial lawyers, are capable of enhancing the state’s capacity for law reform in instances where legal rules are seen as a pathway to economic prosperity. They and other community leaders (typically political and business elites), however, are influenced in their reform efforts by the community economic identities (“CEIs”) they have developed with respect to chosen commercial activities within their jurisdiction. CEIs affect how those who make the crucial development decisions think about the commercial activities and industries in their community (city, state, country or region), and these identities in turn shape what seem like logical choices for that community. This project makes two contributions to the sociology of law and economic development. First, it suggests that the scholarly understanding of “state capacity” should be expanded to include the availability and role of private sector commercial lawyers who can help a state promote and achieve commercial law reforms. Second, the dissertation introduces the concept of CEI into the academic discourse on the relationship between law and economic development. It explores the role commercial lawyers and CEI have played in commercial law reform in the Commonwealth Caribbean post-colonies of Barbados and Jamaica. The institutional similarities between the two countries and the divergent paths they have taken in their approach to commercial law reform make them attractive sites for better understanding the determinants of legal change.
Contributors: Han, Siyang, Nelson, Celeste M, Molecular Biology Department
... Cells in epithelial tissues adhere to an underlying extracellular matrix (ECM) and respond to changes in its physical properties by altering gene expression and behavior. The mechanical properties of the ECM, including the matrix stiffness, are therefore central to developmental, physiological, and pathological processes. In the context of cancer, increased tissue stiffness and interstitial fluid pressure are inherent features of tumorigenesis, which promote malignant transformation of surrounding cells and correlate with poor survival. These processes depend on the synergistic effects of mechanical signals such as ECM stiffness and chemical signals such as Wnt, TGFβ and proteases in the cellular microenvironment. This dissertation explores how mechanical and biochemical signals are integrated to control cell behaviors including proliferation, apoptosis, epithelial-mesenchymal transition (EMT), and invasion. We used engineered substrata to recapitulate the mechanical stiffnesses of the normal mammary gland as well as that of breast tumors. We found that exposure to Wnt3a increased proliferation of mammary epithelial cells cultured on stiff substrata, with compliances characteristic of breast tumors, but not of cells on soft substrata, with compliances comparable to that of normal mammary tissue. Depleting integrin-linked kinase (ILK), which functions as a signaling hub for cellular mechanotransduction, rendered cells unresponsive to Wnt3a on both substrata. Ectopic expression of ILK permitted Wnt3a to induce proliferation of cells on both microenvironments. We further showed that ILK regulates expression of the Wnt receptor frizzled-1 (Fzd1), suggesting the existence of a positive feedback loop between Wnt3a, ILK and Fzd1. Following these findings, we examined the role of ILK and tissue stiffness in regulating cell fate in response to TGFβ1. We found that ILK controls the switch between apoptosis and EMT downstream of TGFβ1 through modulating the fine balance between cell-cell and cell-matrix adhesions. Specifically, depletion of ILK leads to increased E-cadherin-based cell-cell adhesion and cortical actin and decreased focal adhesions and stress fibers in mammary epithelial cells. These changes favor apoptosis and suppress EMT downstream of TGFβ1, regardless of matrix stiffness. Similarly, soft matrix also favors apoptosis downstream of TGFβ1 while stiff matrix promotes EMT, suggesting that depletion of ILK disrupts the ability of a cell to sense a mechanically stiff microenvironment. Taken together, these findings suggest that tissue mechanics regulates the cellular response to Wnt and TGFβ under physiological and pathological microenvironmental conditions. We also highlighted the role of ILK in mediating the responses to matrix stiffness and tuning the proliferative and potential tumorigenic behaviors of mammary epithelial cells to Wnt3a and TGFβ1. These results provide insight into the molecular mechanisms underlying the effects of a pathological microenvironment on neoplastic progression.