Online reinforcement learning (RL) algorithms are often difficult to deploy in complex human-facing applications as they may learn slowly and have poor early performance. To address this, we introduce a practical algorithm for incorporating human insight to speed learning. Our algorithm, Constraint Sampling Reinforcement Learning (CSRL), incorporates prior domain knowledge as constraints/restrictions on the RL policy. It takes in multiple potential policy constraints to maintain robustness to misspecification of individual constraints while leveraging helpful ones to learn quickly. Given a base RL learning algorithm (ex. UCRL, DQN, Rainbow) we propose an upper confidence with elimination scheme that leverages the relationship between the constraints, and their observed performance, to adaptively switch among them. We instantiate our algorithm with DQN-type algorithms and UCRL as base algorithms, and evaluate our algorithm in four environments, including three simulators based on real data: recommendations, educational activity sequencing, and HIV treatment sequencing. In all cases, CSRL learns a good policy faster than baselines.
This paper derives time-uniform confidence sequences (CS) for causal effects in experimental and observational settings. A confidence sequence for a target parameter $\psi$ is a sequence of confidence intervals $(C_t)_{t=1}^\infty$ such that every one of these intervals simultaneously captures $\psi$ with high probability. Such CSs provide valid statistical inference for $\psi$ at arbitrary stopping times, unlike classical fixed-time confidence intervals which require the sample size to be fixed in advance. Existing methods for constructing CSs focus on the nonasymptotic regime where certain assumptions (such as known bounds on the random variables) are imposed, while doubly-robust estimators of causal effects rely on (asymptotic) semiparametric theory. We use sequential versions of central limit theorem arguments to construct large-sample CSs for causal estimands, with a particular focus on the average treatment effect (ATE) under nonparametric conditions. These CSs allow analysts to update statistical inferences about the ATE in lieu of new data, and experiments can be continuously monitored, stopped, or continued for any data-dependent reason, all while controlling the type-I error rate. Finally, we describe how these CSs readily extend to other causal estimands and estimators, providing a new framework for sequential causal inference in a wide array of problems.
In this paper, we introduce a generalization of graphlets to heterogeneous networks called typed graphlets. Informally, typed graphlets are small typed induced subgraphs. Typed graphlets generalize graphlets to rich heterogeneous networks as they explicitly capture the higher-order typed connectivity patterns in such networks. To address this problem, we describe a general framework for counting the occurrences of such typed graphlets. The proposed algorithms leverage a number of combinatorial relationships for different typed graphlets. For each edge, we count a few typed graphlets, and with these counts along with the combinatorial relationships, we obtain the exact counts of the other typed graphlets in o(1) constant time. Notably, the worst-case time complexity of the proposed approach matches the time complexity of the best known untyped algorithm. In addition, the approach lends itself to an efficient lock-free and asynchronous parallel implementation. While there are no existing methods for typed graphlets, there has been some work that focused on computing a different and much simpler notion called colored graphlet. The experiments confirm that our proposed approach is orders of magnitude faster and more space-efficient than methods for computing the simpler notion of colored graphlet. Unlike these methods that take hours on small networks, the proposed approach takes only seconds on large networks with millions of edges. Notably, since typed graphlet is more general than colored graphlet (and untyped graphlets), the counts of various typed graphlets can be combined to obtain the counts of the much simpler notion of colored graphlets. The proposed methods give rise to new opportunities and applications for typed graphlets.
In this work, we reframe the problem of balanced treatment assignment as optimization of a two-sample test between test and control units. Using this lens we provide an assignment algorithm that is optimal with respect to the minimum spanning tree test of Friedman and Rafsky (1979). This assignment to treatment groups may be performed exactly in polynomial time. We provide a probabilistic interpretation of this process in terms of the most probable element of designs drawn from a determinantal point process which admits a probabilistic interpretation of the design. We provide a novel formulation of estimation as transductive inference and show how the tree structures used in design can also be used in an adjustment estimator. We conclude with a simulation study demonstrating the improved efficacy of our method.
Leveraging text, such as social media posts, for causal inferences requires the use of NLP models to 'learn' and adjust for confounders, which could otherwise impart bias. However, evaluating such models is challenging, as ground truth is almost never available. We demonstrate the need for empirical evaluation frameworks for causal inference in natural language by showing that existing, commonly used models regularly disagree with one another on real world tasks. We contribute the first such framework, generalizing several challenges across these real world tasks. Using this framework, we evaluate a large set of commonly used causal inference models based on propensity scores and identify their strengths and weaknesses to inform future improvements. We make all tasks, data, and models public to inform applications and encourage additional research.
In many applied fields, researchers are often interested in tailoring treatments to unit-level characteristics in order to optimize an outcome of interest. Methods for identifying and estimating treatment policies are the subject of the dynamic treatment regime literature. Separately, in many settings the assumption that data are independent and identically distributed does not hold due to inter-subject dependence. The phenomenon where a subject's outcome is dependent on his neighbor's exposure is known as interference. These areas intersect in myriad real-world settings. In this paper we consider the problem of identifying optimal treatment policies in the presence of interference. Using a general representation of interference, via Lauritzen-Wermuth-Freydenburg chain graphs (Lauritzen and Richardson, 2002), we formalize a variety of policy interventions under interference and extend existing identification theory (Tian, 2008; Sherman and Shpitser, 2018). Finally, we illustrate the efficacy of policy maximization under interference in a simulation study.
In this work, we formalize the problem of causal inference over graph-based relational time-series data where each node in the graph has one or more time-series associated to it. We propose causal inference models for this problem that leverage both the graph topology and time-series to accurately estimate local causal effects of nodes. Furthermore, the relational time-series causal inference models are able to estimate local effects for individual nodes by exploiting local node-centric temporal dependencies and topological/structural dependencies. We show that simpler causal models that do not consider the graph topology are recovered as special cases of the proposed relational time-series causal inference model. We describe the conditions under which the resulting estimate can be used to estimate a causal effect, and describe how the Durbin-Wu-Hausman test of specification can be used to test for the consistency of the proposed estimator from data. Empirically, we demonstrate the effectiveness of the causal inference models on both synthetic data with known ground-truth and a large-scale observational relational time-series data set collected from Wikipedia.
In many practical applications of contextual bandits, online learning is infeasible and practitioners must rely on off-policy evaluation (OPE) of logged data collected from prior policies. OPE generally consists of a combination of two components: (i) directly estimating a model of the reward given state and action and (ii) importance sampling. While recent work has made significant advances adaptively combining these two components, less attention has been paid to improving the quality of the importance weights themselves. In this work we present balancing off-policy evaluation (BOP-e), an importance sampling procedure that directly optimizes for balance and can be plugged into any OPE estimator that uses importance sampling. BOP-e directly estimates the importance sampling ratio via a classifier which attempts to distinguish state-action pairs from an observed versus a proposed policy. BOP-e can be applied to continuous, mixed, and multi-valued action spaces without modification and is easily scalable to many observations. Further, we show that minimization of regret in the constructed binary classification problem translates directly into minimizing regret in the off-policy evaluation task. Finally, we provide experimental evidence that BOP-e outperforms inverse propensity weighting-based approaches for offline evaluation of policies in the contextual bandit setting under both discrete and continuous action spaces.
In many practical applications of contextual bandits, online learning is infeasible and practitioners must rely on off-policy evaluation (OPE) of logged data collected from prior policies. OPE generally consists of a combination of two components: (i) directly estimating a model of the reward given state and action and (ii) importance sampling. While recent work has made significant advances adaptively combining these two components, less attention has been paid to improving the quality of the importance weights themselves. In this work we present balancing off-policy evaluation (BOP-e), an importance sampling procedure that directly optimizes for balance and can be plugged into any OPE estimator that uses importance sampling. BOP-e directly estimates the importance sampling ratio via a classifier which attempts to distinguish state-action pairs from an observed versus a proposed policy. BOP-e can be applied to continuous, mixed, and multi-valued action spaces without modification and is easily scalable to many observations. Further, we show that minimization of regret in the constructed binary classification problem translates directly into minimizing regret in the off-policy evaluation task. Finally, we provide experimental evidence that BOP-e outperforms inverse propensity weighting-based approaches for offline evaluation of policies in the contextual bandit setting under both discrete and continuous action spaces.
Many real-world applications give rise to large heterogeneous networks where nodes and edges can be of any arbitrary type (e.g., user, web page, location). Special cases of such heterogeneous graphs include homogeneous graphs, bipartite, k-partite, signed, labeled graphs, among many others. In this work, we generalize the notion of network motifs to heterogeneous networks. In particular, small induced typed subgraphs called typed graphlets (heterogeneous network motifs) are introduced and shown to be the fundamental building blocks of complex heterogeneous networks. Typed graphlets are a powerful generalization of the notion of graphlet (network motif) to heterogeneous networks as they capture both the induced subgraph of interest and the types associated with the nodes in the induced subgraph. To address this problem, we propose a fast, parallel, and space-efficient framework for counting typed graphlets in large networks. We discover the existence of non-trivial combinatorial relationships between lower-order ($k-1$)-node typed graphlets and leverage them for deriving many of the $k$-node typed graphlets in $o(1)$ constant time. Thus, we avoid explicit enumeration of those typed graphlets. Notably, the time complexity matches the best untyped graphlet counting algorithm. The experiments demonstrate the effectiveness of the proposed framework in terms of runtime, space-efficiency, parallel speedup, and scalability as it is able to handle large-scale networks.