Computational Analysis of Intelligent Agents: Social and Strategic Settings

The central motif of this work is prediction and optimization in presence of multiple interacting intelligent agents. We use the phrase `intelligent agents' to imply in some sense, a `bounded rationality', the exact meaning of which varies depending on the setting. Our agents may not be `rational' in the classical game theoretic sense, in that they don't always optimize a global objective. Rather, they rely on heuristics, as is natural for human agents or even software agents operating in the real-world. Within this broad framework we study the problem of influence maximization in social networks where behavior of agents is myopic, but complication stems from the structure of interaction networks. In this setting, we generalize two well-known models and give new algorithms and hardness results for our models. Then we move on to models where the agents reason strategically but are faced with considerable uncertainty. For such games, we give a new solution concept and analyze a real-world game using out techniques. Finally, the richest model we consider is that of Network Cournot Competition which deals with strategic resource allocation in hypergraphs, where agents reason strategically and their interaction is specified indirectly via player's utility functions. For this model, we give the first equilibrium computability results. In all of the above problems, we assume that payoffs for the agents are known. However, for real-world games, getting the payoffs can be quite challenging. To this end, we also study the inverse problem of inferring payoffs, given game history. We propose and evaluate a data analytic framework and we show that it is fast and performant.

The Relation between Substance Use and Medication Adherence among HIV Positive Substance Users in Residential Treatment

Poor medication adherence is problematic among HIV positive, low-income African-American substance users. Substance use has been shown to be associated with poor medication adherence, though we do not know the mechanism that underlies this relationship. Lack of positive environmental rewards and the propensity to discount delayed rewards may be possible mechanisms to explain this relationship. Using baseline data from a randomized controlled trial, we examined the relationships between substance use and medication adherence, testing both environmental rewards and delay discounting as independent mediators. There was a main effect of substance use on adherence, such that high frequency of substance use predicted poor adherence. There was also a main effect of environmental rewards on adherence, such that a lack of environmental reinforcement predicted poor adherence. This study shed light on the processes that contribute to low adherence, namely substance use and lack of environmental contingencies, and suggests important targets for intervention.

DIFFERENCES IN PSYCHOPATHOLOGY AMONG SUBSTANCE USERS IN RESIDENTIAL TREATMENT RELATED TO HIV STATUS

HIV-positive individuals engage in substance use at higher rates than the general population and are more likely to also suffer from concurrent psychiatric disorders and substance use disorders. Despite this, little is known about the unique clinical concerns of HIV-positive individuals entering substance use treatment. This study examined the clinical characteristics of clients (N=1712) entering residential substance use treatment as a function of self-reported HIV status (8.65% HIV-positive). Results showed higher levels of concurrent substance use and psychiatric disorders for HIV-positive individuals, who were also significantly more likely to meet criteria for bipolar disorder and borderline personality disorder. Past diagnoses of depression, posttraumatic stress disorder, and social phobia were also significantly more common. Study findings indicate a need to provide more intensive care for HIV-positive individuals, including resources targeted at concurrent psychiatric problems, to ensure positive treatment outcomes following residential substance use treatment discharge.

Efficient Secure Computation for Real-world Settings and Security Models

Secure computation involves multiple parties computing a common function while keeping their inputs private, and is a growing field of cryptography due to its potential for maintaining privacy guarantees in real-world applications. However, current secure computation protocols are not yet efficient enough to be used in practice. We argue that this is due to much of the research effort being focused on generality rather than specificity. Namely, current research tends to focus on constructing and improving protocols for the strongest notions of security or for an arbitrary number of parties. However, in real-world deployments, these security notions are often too strong, or the number of parties running a protocol would be smaller. In this thesis we make several steps towards bridging the efficiency gap of secure computation by focusing on constructing efficient protocols for specific real-world settings and security models. In particular, we make the following four contributions: - We show an efficient (when amortized over multiple runs) maliciously secure two-party secure computation (2PC) protocol in the multiple-execution setting, where the same function is computed multiple times by the same pair of parties. - We improve the efficiency of 2PC protocols in the publicly verifiable covert security model, where a party can cheat with some probability but if it gets caught then the honest party obtains a certificate proving that the given party cheated. - We show how to optimize existing 2PC protocols when the function to be computed includes predicate checks on its inputs. - We demonstrate an efficient maliciously secure protocol in the three-party setting.

ESSAYS ON DISPLACED WORKERS AND RESIDENTIAL MIGRATION

In this dissertation, I explore how workers’ human capital, local industry composition, and business cycles affect employment outcomes and residential migration for job losers and other workers. I first examine whether the poor employment outcomes of job losers are due to a lack of jobs that require their human capital within their local labor market. I answer this question by analyzing the extent to which the industry composition in the job loser’s local labor market affects employment outcomes when job loss occurs during expansions and during recessions. I find that if job losers reside in an area with a high employment concentration of their original industry of employment, they are 2.1-2.8 percent more likely to be re-employed at another job if job loss occurs during an expansion; I find an insignificant relationship in most specifications when job loss occurs during a recession, and in some specifications I even find a negative relationship between industry concentration and employment. I conclude that the industry composition within an area matters for job losers, since firms are more willing to hire workers from within their own industry, as these workers have more relevant accumulated human capital. However, firms are less likely to hire during a recession, making job losers’ human capital less important for job finding. Next, Erika McEntarfer, Henry Hyatt, and I examine whether the business cycle affects earnings changes for job losers, and the factors that explain these differences across time. We find that job losers who lost their job during the Great Recession have earnings changes that are 10 percent more negative relative to other job losers from other periods. This result is driven primarily by longer nonemployment lengths and worse subsequent job matches. Finally, Erika McEntarfer, Henry Hyatt, Alexandria Zhang, and I explore the extent to which residential migration is driven by job opportunities. We use four databases and find that changes in job moves explain some of the changes in residential migration, but the relationship is not as strong as previously documented. We find that migration patterns differ across databases, with some databases documenting steeper declines and more cyclicality.