48670 research outputs found
Sort by
Beyond Statistical Fairness
No full textIn recent years, a great deal of fairness notions has been proposed. Yet, most of them take a reductionist approach by indirectly viewing fairness as equalizing some error statistic across pre-defined groups. This thesis aims to explore some ideas as to how to go beyond such statistical fairness frameworks. First, we consider settings in which the right notion of fairness may not be captured by simple mathematical definitions but might be more complex and nuanced and thus require elicitation from individual or collective stakeholders. By asking stakeholders to make pairwise comparisons to learn which pair of individuals should be treated similarly, we show how to approximately learn the most accurate classifier or converge to such one subject to the elicited fairness constraints. We consider an offline setting where the pairwise comparisons must be made prior to training a model and an online setting where one can continually provide fairness feedback to the deployed model in each round. We also report preliminary findings of a behavioral study of our framework using human-subject fairness constraints elicited on the COMPAS criminal recidivism dataset. Second, unlike most of the statistical fairness framework that promises fairness for pre-defined and often coarse groups, we provide fairness guarantees for finer subgroups, such as all possible intersections of the pre-defined groups, in the context of uncertainty estimation in both offline and online setting. Our framework gives uncertainty guarantees that are more locally sensible than the ones given by conformal prediction techniques; our uncertainty estimates are valid even when averaged over any subgroup, but uncertainty estimates in conformal predictions are usually only valid when averaged over the entire population
Essays on Corporate Political Strategies and Firm-Stakeholder Relationships
No full textThis dissertation examines the nonmarket strategies that firms use to respond to and to create change in the social and political environment. In the first chapter, I document the extent of corporate political connections across Europe and analyze the institutional contingencies of their impact. I find that a political connection increases the annual value of public procurement contracts awarded to firms, but that this positive impact is moderated by accountability institutions that enable different bodies of government to check one another’s power. When accountability institutions are weaker, firms become more likely to hold political connections, and the competition between rival firms to form political connections intensifies, suggesting that nonmarket competition in the political arena becomes increasingly consequential as the constraints of political patronage are weakened. In the second chapter, I analyze the grantmaking behavior of private foundations in the United States and find evidence consistent with philanthropy being used as a tax-advantaged method to influence politics through grantmaking to politically active nonprofits. In the third chapter, co-authored with Brian Ganson and Witold J. Henisz, we develop theories to explain how a firm’s relational strategies might impact the structure of relationships, and hence conflict risk, between groups in conflict-affected areas. Firms can affect broader societal outcomes when their actions, such as the unequal distribution of benefits to majority groups over minority groups, change the tenor of conflict and cooperation between identity groups delineated by ethnic, class, geographic, cultural, or other divides. As such, firms become implicated in shaping how they and their stakeholders reach mutually acceptable settlements in response to collective challenges. In the fourth chapter, co-authored with Anne S. Jamison, we propose a set of a methods to map the structure of relationships between firms and stakeholders in conflict-affected areas. We utilize geo- and time-tagged data from media reports and apply natural language processing techniques to construct a network of relationships between political figures, civil society actors, and firms in Rwanda
Practical Network Programming Automation
No full textNetwork configurations are notoriously hard to write and maintain correctly. It requires expertise about the domain to write, frequent and laborious updates, and sometimes formal proof to ensure the absence of certain mistakes. The problem becomes more challenging with the popularity of software-defined network(SDN) in recent years, which aims to give users more flexible control over the network’s dynamic behaviors. There has been research on automating the process of configuring the network. However, much of it requires users to learn a specific programming abstraction or interface. Since network operators are a group generally unfamiliar with programming, using these systems may go beyond their abilities. It is also hard to ensure these systems are scalable and accurate enough for real-world usecases. They mostly lack both design considerations to address scalability and accuracy, and also a systematic evaluation of the two metrics in practical scenarios. In this work, we propose a series of approaches to automate network programming. They are based on specifications that are easy and natural to obtain by network operators. We also apply novel program analysis techniques to speed up the process of finding a program that can accurately capture the intention of the specification. We have evaluated our systems on a broad range of benchmarks obtained from real-world data. They have shown ability to finish complex programming tasks within minutes and achieved very high accuracy
Inference of Shared Genetic Architecture with Genome-Wide Association Data
No full textGenome-wide association studies (GWAS) have identified thousands of genetic variants associated with complex traits. Many complex traits and diseases share common genetic architecture. Studying the shared genetic architecture provides valuable insights into the underlying disease mechanisms. In this dissertation, we develop several statistical methods for investigating the shared genetic architecture based on GWAS data. We first discuss the quantification and estimation of the shared genetic architecture based on genetic covariance, which is defined as the underlying covariance of the genetic effects. We develop a unified approach to robust estimation and inference for genetic covariance of general outcomes that can be associated with genetic variants nonlinearly. The theoretical analysis shows that the proposed estimator is robust under certain model mis-specification. Various numerical experiments are performed to support the theoretical results. Application of this method to an outbred mice GWAS data set reveals interesting genetic covariance among different mice developmental traits. We then consider a practical challenge when the raw genotype data are unavailable, but only the GWAS summary association statistics are available. We develop a method of moments estimator of genetic correlation between two traits in the framework of high dimensional linear models. Theoretical properties of the estimator in terms of consistency and asymptotic normality are provided. Simulations and real data analysis results show that the proposed estimator is more robust and has better interpretability than the LD score regression method under different genetic architectures. Finally, in chapter 4 we discuss the problem of genome-wide detection and identification of shared genetic association, which is a global assessment of the existence of shared genetic architecture. The challenge is that the linkage disequilibrium (LD) between the SNPs makes test statistics highly dependent, which complicates the detection and identification. To account for such a dependency, an eigenvector-projected score statistic is proposed and a max-type test statistic (max-block) is developed for the genome-wide detection of shared associations. The max-block method is easy to calculate and is shown to control the genome-wide error rate. The method is applied to study shared cross-trait associations in 10 pediatric autoimmune diseases, leading to several regions that explain the genetic sharing between diseases
Essays on the Use of A/B Testing among e-Commerce Practitioners
No full textRandomized experiments—often called A/B tests in industrial settings—are an increasingly important element in the management of many organizations. While some firms have long had both the managerial and technical know-how to use experiments for making key decisions, new forms of software and internet infrastructure have dramatically lowered the cost of conducting A/B tests online, opening up the practice to an entirely new set of organizations. This dissertation studies the practice of A/B testing among this new wave of practitioners, characterized primarily as e-commerce businesses that have adopted new forms of low cost, easy to use, third-party experimentation software. The first two chapters of this document study A/B testing as its own distinct phenomena in digital business, answering questions about the prevalence of p-hacking among e-commerce practitioners and the nature of how firms use A/B testing software in the real world. The final chapter demonstrates how e-commerce firms can use A/B tests and recent developments in causal machine learning for improved customer targeting and price discrimination. As a whole, this work demonstrates the growing importance of A/B testing and causal reasoning as a key factor in the future of managerial decision making
How Gloomy is the Retirement Outlook for Millennials?
Full text linkSocial, economic, demographic, and public policy shifts have made Millennial retirement security a pressing concern. Many recent trends threaten financial security for future generations of retirees. Male labor force participation pre-age 55 has slumped, men’s median earnings have stagnated, marriage and homeownership rates are falling, debt levels remain high, and out-of-pocket spending on medical and long-term services and supports are rising. Other trends are more encouraging, such as women’s higher earnings, the rise in labor force participation at older ages, and improvements in educational attainment. We use a dynamic microsimulation model to project how various forces might play out over the next 30 years to shape the retirement security of US residents born in the 1980s. Our projections show that median age-70 income will be higher for Millennials than previous generations, but this cohort faces a higher risk of seeing falling living standards in retirement
Communitas: Building Community through Leisure and Collective Joy
Full text linkRelationships and belonging are fundamental needs for well-being. At the same time, our culture is becoming increasingly individualistic and loneliness is pervasive. Our traditional community and relational frameworks, such as religious institutions and associations, are also diminishing. This capstone explores leisure as a possible pathway for building social capital and community well-being. We will look at why leisure activities based in recreation, play and the humanities have the capacity to build community and enduring connections with others, while also having individual well-being benefits. This capstone features some exemplary groups that utilize leisure activities and have built strong communities through positive psychology constructs such as hive psychology, self-efficacy, collective effervescence, mattering, resilience and belonging. We will explore opportunities to overlap the fields of positive psychology and leisure in both research and application. We will remind ourselves that joy and belonging are powerful forces and even more powerful when they come together
Collation Model for Oversize LJS 254: [Buch von den probierten Künsten]
No full textIllustrated treatise on gunpowder and artillery; probably a copy of an early recension of Franz Helm\u27s work Buch von den probierten Kü̈̈nsten that first appeared between 1527 and 1535. The title page includes language similar to that title and a date of 1562 (f. ii). The date is echoed in a note added at the end of the manuscript in another hand; this note also attributes the copying of the manuscript to Georg Ganser of Cham and suggests that it was translated from French into German (f. 265r).https://repository.upenn.edu/sims_models/1053/thumbnail.jp
Collation Model for LJS 451: [Commentary on the Doctrinale puerorum]
Full text linkAnonymous commentary on a 12th-century didactic poem in Latin, composed by the French Franciscan Alexander de Villa Dei addressing parts of speech, syntax, and rhetorical figures. Opening of a letter by Albert II, Duke of Austria, Styria, and Carinthia, transcribed in German by a later hand on last verso.https://repository.upenn.edu/sims_models/1057/thumbnail.jp
Collation Model for LJS 304: [Dogale issued by Doge Alvise Mocenigo to Girolamo Pruli, podestà of Brescia].
Full text linkCommission from Alvise Mocenigo, doge of Venice, addressed to Girolamo Priuli (not to be confused with the Girolamo Priuli, of the same family, who served as doge of Venice and died in 1567), concerning Priuli\u27s duties, rights, and obligations while holding the position of podestà (civil administrator) of Brescia, a mainland community west of Venice under Venetian control.https://repository.upenn.edu/sims_models/1077/thumbnail.jp