Peer review is fundamental to academic research, yet its effectiveness hinges on reviewers voluntarily exerting costly effort to evaluate submissions. Current mechanisms used by conference organizers, such as random or affinity-based matching, fail to adequately incentivize effort, leading to suboptimal outcomes. This study investigates how peer review markets can be designed to maximize the acceptance of high-quality submissions while minimizing the acceptance of low-quality ones.

We study effect propagation in a causal directed acyclic graph (DAG), with the goal of providing a flow-based decomposition of the effect (i.e., change in the outcome variable) as a result of changes in the source variables. We first compare various ideas on causality to quantify effect propagation, such as direct and indirect effects, path-specific effects, and degree of responsibility. We discuss the shortcomings of such approaches and propose a flow-based methodology, which we call recursive Shapley value (RSV). By considering a broader set of counterfactuals than existing methods, RSV obeys a unique adherence to four desirable flow-based axioms. Further, we provide a general path-based characterization of RSV for an arbitrary non-parametric structural equations model (SEM) defined on the underlying DAG. Interestingly, for the special class of linear SEMs, RSV exhibits a simple and tractable characterization (and hence, computation), which recovers the classical method of path coefficients and is equivalent to path-specific effects. For non-parametric SEMs, we use our general characterization to develop an unbiased Monte-Carlo estimation procedure with an exponentially decaying sample complexity. We showcase the application of RSV on two challenging problems on causality (causal overdetermination and causal unfairness).

We provide an optimization-based framework to perform counterfactual analysis in a dynamic model with hidden states. Our framework is grounded in the “abduction, action, and prediction” approach to answer counterfactual queries and handles two key challenges where (1) the states are hidden and (2) the model is dynamic. Recognizing the lack of knowledge on the underlying causal mechanism and the possibility of infinitely many such mechanisms, we optimize over this space and compute upper and lower bounds on the counterfactual quantity of interest. Our work brings together ideas from causality, state-space models, simulation, and optimization, and we apply it on a breast cancer case study. To the best of our knowledge, we are the first to compute lower and upper bounds on a counterfactual query in a dynamic latent-state model.

We study the attribution problem in a graphical model, wherein the objective is to quantify how the effect of changes at the source nodes propagates through the graph. We develop a model-agnostic flow-based attribution method, called recursive Shapley value (RSV). RSV generalizes a number of existing node-based methods and uniquely satisfies a set of flow-based axioms. In addition to admitting a natural characterization for linear models and facilitating mediation analysis for non-linear models, RSV satisfies a mix of desirable properties discussed in the recent literature, including implementation invariance, sensitivity, monotonicity, and affine scale invariance.

One of the central challenges in online advertising is attribution, namely, assessing the contribution of individual advertiser actions including emails, display ads and search ads to eventual conversion. Several heuristics are used for attribution in practice; however, there is no formal justification for them and many of these fail even in simple canonical settings. The main contribution in this work is to develop an axiomatic framework for attribution in online advertising. In particular, we consider a Markovian model for the user journey through the conversion funnel, in which ad actions may have disparate impacts at different stages. We propose a novel attribution metric, that we refer to as counterfactual adjusted Shapley value, which inherits the desirable properties of the traditional Shapley value. Furthermore, we establish that this metric coincides with an adjusted “unique-uniform” attribution scheme. This scheme is efficiently computable and implementable and can be interpreted as a correction to the commonly used uniform attribution scheme.

Temporal difference learning (TD) is a simple iterative algorithm used to estimate the value function corresponding to a given policy in a Markov decision process. Although TD is one of the most widely used algorithms in reinforcement learning, its theoretical analysis has proved challenging and few guarantees on its statistical efficiency are available. In this work, we provide a \emph{simple and explicit finite time analysis} of temporal difference learning with linear function approximation. Except for a few key insights, our analysis mirrors standard techniques for analyzing stochastic gradient descent algorithms, and therefore inherits the simplicity and elegance of that literature. A final section of the paper shows that all of our main results extend to the study of a variant of Q-learning applied to optimal stopping problems.