Search for: All records

Deep reinforcement learning has demonstrated re- markable achievements across diverse domains such as video games, robotic control, autonomous driving, and drug discovery. Common methodologies in partially observable domains largely lean on end-to-end learning from high-dimensional observations, such as images, without explicitly reasoning about true state. We suggest an alternative direction, introducing the Partially Supervised Reinforcement Learning (PSRL) framework. At the heart of PSRL is the fusion of both supervised and unsupervised learning. The approach leverages a state estimator to distill supervised semantic state information from high-dimensional observations which are often fully observable at training time. This yields more interpretable policies that compose state predictions with control. In parallel, it captures an unsupervised latent representation. These two—the semantic state and the latent state—are then fused and utilized as inputs to a policy network. This juxtaposition offers practitioners a flexible and dynamic spectrum: from emphasizing supervised state information to integrating richer, latent insights. Extensive experimental results indicate that by merging these dual representations, PSRL offers a balance, enhancing interpretability while preserving, and often significantly outperforming, the performance benchmarks set by traditional methods in terms of reward and convergence speed. 

Adversarial machine learning (AML) research is concerned with robustness of machine learning models and algorithms to malicious tampering. Originating at the intersection between machine learning and cybersecurity, AML has come to have broader research appeal, stretching traditional notions of security to include applications of computer vision, natural language processing, and network science. In addition, the problems of strategic classification, algorithmic recourse, and counterfactual explanations have essentially the same core mathematical structure as AML, despite distinct motivations. I give a simplified overview of the central problems in AML, and then discuss both the security-motivated AML domains, and the problems above unrelated to security. These together span a number of important AI subdisciplines, but can all broadly be viewed as concerned with trustworthy AI. My goal is to clarify both the technical connections among these, as well as the substantive differences, suggesting directions for future research. 

Social networks arise as a result of complex interactions among people, and homophily plays an important role in this process. If we view homophily as a dominant force in network formation and associate each node with a collection of features, this process gives rise to spatial networks, with likelihood of an edge an increasing function of feature similarity among its incident nodes. A link prediction problem in such spatial networks then amounts to deter- mining whether the pair of nodes are sufficiently close according to this edge likelihood function. We undertake the first algorithmic study of the adversarial side of this problem in which the adversary manipulates features of a subset of nodes on the network to pre- vent predicting target edges. We show that this problem is NP-hard, even if the edge likelihood function is convex. On the other hand, if this function is convex, we show that the problem can be solved with convex programming when the set of nodes that the adversary needs to manipulate is fixed. Furthermore, if the edge likelihood function is linear, we present approximation algorithms for the case when the features are binary, and we wish to hide only a single edge, and for the case when the features are real-valued but we need to hide an arbitrary collection of edges. 

The increasing automation of high-stakes decisions with direct impact on the lives and well-being of individuals raises a number of important considerations. Prominent among these is strategic behavior by individuals hoping to achieve a more desirable outcome. Two forms of such behavior are commonly studied: 1) misreporting of individual attributes, and 2) recourse, or actions that truly change such attributes. The former involves deception, and is inherently undesirable, whereas the latter may well be a desirable goal insofar as it changes true individual qualification. We study misreporting and recourse as strategic choices by individuals within a unified framework. In particular, we propose auditing as a means to incentivize recourse actions over attribute manipulation, and characterize optimal audit policies for two types of principals, utility-maximizing and recourse-maximizing. Additionally, we consider subsidies as an incentive for recourse over manipulation, and show that even a utility-maximizing principal would be willing to devote a considerable amount of audit budget to providing such subsidies. Finally, we consider the problem of optimizing fines for failed audits, and bound the total cost incurred by the population as a result of audits.