arXiv subjects

William Zame

Publications and source records attributed to William Zame.

8 recordsLinked to original sources

arXiv2021

SDF-Bayes: Cautious Optimism in Safe Dose-Finding Clinical Trials with Drug Combinations and Heterogeneous Patient Groups

Phase I clinical trials are designed to test the safety (non-toxicity) of drugs and find the maximum tolerated dose (MTD). This task becomes significantly more challenging when multiple-drug dose-combinations (DC) are involved, due to the inherent conflict between the exponentially increasing DC candidates and the limited patient budget. This paper proposes a novel Bayesian design, SDF-Bayes, for finding the MTD for drug combinations in the presence of safety constraints. Rather than the conventional principle of escalating or de-escalating the current dose of one drug (perhaps alternating between drugs), SDF-Bayes proceeds by cautious optimism: it chooses the next DC that, on the basis of current information, is most likely to be the MTD (optimism), subject to the constraint that it only chooses DCs that have a high probability of being safe (caution). We also propose an extension, SDF-Bayes-AR, that accounts for patient heterogeneity and enables heterogeneous patient recruitment. Extensive experiments based on both synthetic and real-world datasets demonstrate the advantages of SDF-Bayes over state of the art DC trial designs in terms of accuracy and safety.

cs.LG↗

arXiv2020

Learning outside the Black-Box: The pursuit of interpretable models

Machine Learning has proved its ability to produce accurate models but the deployment of these models outside the machine learning community has been hindered by the difficulties of interpreting these models. This paper proposes an algorithm that produces a continuous global interpretation of any given continuous black-box function. Our algorithm employs a variation of projection pursuit in which the ridge functions are chosen to be Meijer G-functions, rather than the usual polynomial splines. Because Meijer G-functions are differentiable in their parameters, we can tune the parameters of the representation by gradient descent; as a consequence, our algorithm is efficient. Using five familiar data sets from the UCI repository and two familiar machine learning algorithms, we demonstrate that our algorithm produces global interpretations that are both highly accurate and parsimonious (involve a small number of terms). Our interpretations permit easy understanding of the relative importance of features and feature interactions. Our interpretation algorithm represents a leap forward from the previous state of the art.

cs.LG↗

arXiv2020

Robust Recursive Partitioning for Heterogeneous Treatment Effects with Uncertainty Quantification

Subgroup analysis of treatment effects plays an important role in applications from medicine to public policy to recommender systems. It allows physicians (for example) to identify groups of patients for whom a given drug or treatment is likely to be effective and groups of patients for which it is not. Most of the current methods of subgroup analysis begin with a particular algorithm for estimating individualized treatment effects (ITE) and identify subgroups by maximizing the difference across subgroups of the average treatment effect in each subgroup. These approaches have several weaknesses: they rely on a particular algorithm for estimating ITE, they ignore (in)homogeneity within identified subgroups, and they do not produce good confidence estimates. This paper develops a new method for subgroup analysis, R2P, that addresses all these weaknesses. R2P uses an arbitrary, exogenously prescribed algorithm for estimating ITE and quantifies the uncertainty of the ITE estimation, using a construction that is more robust than other methods. Experiments using synthetic and semi-synthetic datasets (based on real data) demonstrate that R2P constructs partitions that are simultaneously more homogeneous within groups and more heterogeneous across groups than the partitions produced by other methods. Moreover, because R2P can employ any ITE estimator, it also produces much narrower confidence intervals with a prescribed coverage guarantee than other methods.

stat.ML↗

arXiv2020

AutoCP: Automated Pipelines for Accurate Prediction Intervals

Successful application of machine learning models to real-world prediction problems, e.g. financial forecasting and personalized medicine, has proved to be challenging, because such settings require limiting and quantifying the uncertainty in the model predictions, i.e. providing valid and accurate prediction intervals. Conformal Prediction is a distribution-free approach to construct valid prediction intervals in finite samples. However, the prediction intervals constructed by Conformal Prediction are often (because of over-fitting, inappropriate measures of nonconformity, or other issues) overly conservative and hence inadequate for the application(s) at hand. This paper proposes an AutoML framework called Automatic Machine Learning for Conformal Prediction (AutoCP). Unlike the familiar AutoML frameworks that attempt to select the best prediction model, AutoCP constructs prediction intervals that achieve the user-specified target coverage rate while optimizing the interval length to be accurate and less conservative. We tested AutoCP on a variety of datasets and found that it significantly outperforms benchmark algorithms.

cs.LG↗

arXiv2019

Optimal Piecewise Local-Linear Approximations

Existing works on "black-box" model interpretation use local-linear approximations to explain the predictions made for each data instance in terms of the importance assigned to the different features for arriving at the prediction. These works provide instancewise explanations and thus give a local view of the model. To be able to trust the model it is important to understand the global model behavior and there are relatively fewer works which do the same. Piecewise local-linear models provide a natural way to extend local-linear models to explain the global behavior of the model. In this work, we provide a dynamic programming based framework to obtain piecewise approximations of the black-box model. We also provide provable fidelity, i.e., how well the explanations reflect the black-box model, guarantees. We carry out simulations on synthetic and real datasets to show the utility of the proposed approach. At the end, we show that the ideas developed for our framework can also be used to address the problem of clustering for one-dimensional data. We give a polynomial time algorithm and prove that it achieves optimal clustering.

cs.LG↗

arXiv2013

Designing Efficient Resource Sharing For Impatient Players Using Limited Monitoring

The problem of efficient sharing of a resource is nearly ubiquitous. Except for pure public goods, each agent's use creates a negative externality; often the negative externality is so strong that efficient sharing is impossible in the short run. We show that, paradoxically, the impossibility of efficient sharing in the short run enhances the possibility of efficient sharing in the long run, even if outcomes depend stochastically on actions, monitoring is limited and users are not patient. We base our analysis on the familiar framework of repeated games with imperfect public monitoring, but we extend the framework to view the monitoring structure as chosen by a designer who balances the benefits and costs of more accurate observations and reports. Our conclusions are much stronger than in the usual folk theorems: we do not require a rich signal structure or patient users and provide an explicit online construction of equilibrium strategies.

cs.GT↗

arXiv2012

Designing Practical Distributed Exchange for Online Communities

In many online systems, individuals provide services for each other; the recipient of the service obtains a benefit but the provider of the service incurs a cost. If benefit exceeds cost, provision of the service increases social welfare and should therefore be encouraged -- but the individuals providing the service gain no (immediate) benefit from providing the service and hence have an incentive to withhold service. Hence there is scope for designing a system that improves welfare by encouraging exchange. To operate successfully within the confines of the online environment, such a system should be distributed, practicable, and consistent with individual incentives. This paper proposes and analyzes a simple such system that relies on the exchange of {\em tokens}; the emphasis is on the design of a protocol (number of tokens and suggested strategies). We provide estimates for the efficiency of such protocols and show that choosing the right protocol will lead to almost full efficiency if agents are sufficiently patient. However, choosing the wrong protocols may lead to an enormous loss of efficiency.

cs.GT↗

arXiv2012

Designing Information Revelation and Intervention with an Application to Flow Control

There are many familiar situations in which a manager seeks to design a system in which users share a resource, but outcomes depend on the information held and actions taken by users. If communication is possible, the manager can ask users to report their private information and then, using this information, instruct them on what actions they should take. If the users are compliant, this reduces the manager's optimization problem to a well-studied problem of optimal control. However, if the users are self-interested and not compliant, the problem is much more complicated: when asked to report their private information, the users might lie; upon receiving instructions, the users might disobey. Here we ask whether the manager can design the system to get around both of these difficulties. To do so, the manager must provide for the users the incentives to report truthfully and to follow the instructions, despite the fact that the users are self-interested. For a class of environments that includes many resource allocation games in communication networks, we provide tools for the manager to design an efficient system. In addition to reports and recommendations, the design we employ allows the manager to intervene in the system after the users take actions. In an abstracted environment, we find conditions under which the manager can achieve the same outcome it could if users were compliant, and conditions under which it does not. We then apply our framework and results to design a flow control management system.

cs.GT↗