We propose a novel nonparametric sequential test for composite hypotheses for means of multiple data streams. Our proposed method, \emph{peeking with expectation-based averaged capital} (PEAK), builds upon the testing-as-betting framework and provides a non-asymptotic $\alpha$-level test across any stopping time. PEAK is computationally tractable and efficiently rejects hypotheses that are incorrect across all potential distributions that satisfy our nonparametric assumption, enabling joint composite hypothesis testing on multiple streams of data. We numerically validate our theoretical findings under the best arm identification and threshold identification in the bandit setting, illustrating both the competitive performance and the computational efficiency of our method against state-of-the-art testing methods.
In digital health, the strategy of allocating a limited treatment budget across available risk times is crucial to reduce user fatigue. This strategy, however, encounters a significant obstacle due to the unknown actual number of risk times, a factor not adequately addressed by existing methods lacking theoretical guarantees. This paper introduces, for the first time, the online uniform risk times sampling problem within the approximation algorithm framework. We propose two online approximation algorithms for this problem, one with and one without learning augmentation, and provide rigorous theoretical performance guarantees for them using competitive ratio analysis. We assess the performance of our algorithms using both synthetic experiments and a real-world case study on HeartSteps mobile applications.
This work addresses the problem of automated covariate selection under limited prior knowledge. Given an exposure-outcome pair {X,Y} and a variable set Z of unknown causal structure, the Local Discovery by Partitioning (LDP) algorithm partitions Z into subsets defined by their relation to {X,Y}. We enumerate eight exhaustive and mutually exclusive partitions of any arbitrary Z and leverage this taxonomy to differentiate confounders from other variable types. LDP is motivated by valid adjustment set identification, but avoids the pretreatment assumption commonly made by automated covariate selection methods. We provide theoretical guarantees that LDP returns a valid adjustment set for any Z that meets sufficient graphical conditions. Under stronger conditions, we prove that partition labels are asymptotically correct. Total independence tests is worst-case quadratic in |Z|, with sub-quadratic runtimes observed empirically. We numerically validate our theoretical guarantees on synthetic and semi-synthetic graphs. Adjustment sets from LDP yield less biased and more precise average treatment effect estimates than baselines, with LDP outperforming on confounder recall, test count, and runtime for valid adjustment set discovery.
We consider the problem of estimating a scalar target parameter in the presence of nuisance parameters. Replacing the unknown nuisance parameter with a nonparametric estimator, e.g.,a machine learning (ML) model, is convenient but has shown to be inefficient due to large biases. Modern methods, such as the targeted minimum loss-based estimation (TMLE) and double machine learning (DML), achieve optimal performance under flexible assumptions by harnessing ML estimates while mitigating the plug-in bias. To avoid a sub-optimal bias-variance trade-off, these methods perform a debiasing step of the plug-in pre-estimate. Existing debiasing methods require the influence function of the target parameter as input. However, deriving the IF requires specialized expertise and thus obstructs the adaptation of these methods by practitioners. We propose a novel way to debias plug-in estimators which (i) is efficient, (ii) does not require the IF to be implemented, (iii) is computationally tractable, and therefore can be readily adapted to new estimation problems and automated without analytic derivations by the user. We build on the TMLE framework and update a plug-in estimate with a regularized likelihood maximization step over a nonparametric model constructed with a reproducing kernel Hilbert space (RKHS), producing an efficient plug-in estimate for any regular target parameter. Our method, thus, offers the efficiency of competing debiasing techniques without sacrificing the utility of the plug-in approach.
Contextual bandit algorithms are commonly used in digital health to recommend personalized treatments. However, to ensure the effectiveness of the treatments, patients are often requested to take actions that have no immediate benefit to them, which we refer to as pro-treatment actions. In practice, clinicians have a limited budget to encourage patients to take these actions and collect additional information. We introduce a novel optimization and learning algorithm to address this problem. This algorithm effectively combines the strengths of two algorithmic approaches in a seamless manner, including 1) an online primal-dual algorithm for deciding the optimal timing to reach out to patients, and 2) a contextual bandit learning algorithm to deliver personalized treatment to the patient. We prove that this algorithm admits a sub-linear regret bound. We illustrate the usefulness of this algorithm on both synthetic and real-world data.
In the problem of active sequential hypotheses testing (ASHT), a learner seeks to identify the true hypothesis $h^*$ from among a set of hypotheses $H$. The learner is given a set of actions and knows the outcome distribution of any action under any true hypothesis. While repeatedly playing the entire set of actions suffices to identify $h^*$, a cost is incurred with each action. Thus, given a target error $\delta>0$, the goal is to find the minimal cost policy for sequentially selecting actions that identify $h^*$ with probability at least $1 - \delta$. This paper provides the first approximation algorithms for ASHT, under two types of adaptivity. First, a policy is partially adaptive if it fixes a sequence of actions in advance and adaptively decides when to terminate and what hypothesis to return. Under partial adaptivity, we provide an $O\big(s^{-1}(1+\log_{1/\delta}|H|)\log (s^{-1}|H| \log |H|)\big)$-approximation algorithm, where $s$ is a natural separation parameter between the hypotheses. Second, a policy is fully adaptive if action selection is allowed to depend on previous outcomes. Under full adaptivity, we provide an $O(s^{-1}\log (|H|/\delta)\log |H|)$-approximation algorithm. We numerically investigate the performance of our algorithms using both synthetic and real-world data, showing that our algorithms outperform a previously proposed heuristic policy.
Given only data generated by a standard confounding graph with unobserved confounder, the Average Treatment Effect (ATE) is not identifiable. To estimate the ATE, a practitioner must then either (a) collect deconfounded data; (b) run a clinical trial; or (c) elucidate further properties of the causal graph that might render the ATE identifiable. In this paper, we consider the benefit of incorporating a (large) confounded observational dataset alongside a (small) deconfounded observational dataset when estimating the ATE. Our theoretical results show that the inclusion of confounded data can significantly reduce the quantity of deconfounded data required to estimate the ATE to within a desired accuracy level. Moreover, in some cases---say, genetics---we could imagine retrospectively selecting samples to deconfound. We demonstrate that by strategically selecting these examples based upon the (already observed) treatment and outcome, we can reduce our data dependence further. Our theoretical and empirical results establish that the worst-case relative performance of our approach (vs. a natural benchmark) is bounded while our best-case gains are unbounded. Next, we demonstrate the benefits of selective deconfounding using a large real-world dataset related to genetic mutation in cancer. Finally, we introduce an online version of the problem, proposing two adaptive heuristics.