We introduce an adaptive method with formal quality guarantees for weak supervision in a non-stationary setting. Our goal is to infer the unknown labels of a sequence of data by using weak supervision sources that provide independent noisy signals of the correct classification for each data point. This setting includes crowdsourcing and programmatic weak supervision. We focus on the non-stationary case, where the accuracy of the weak supervision sources can drift over time, e.g., because of changes in the underlying data distribution. Due to the drift, older data could provide misleading information to infer the label of the current data point. Previous work relied on a priori assumptions on the magnitude of the drift to decide how much data to use from the past. In contrast, our algorithm does not require any assumptions on the drift, and it adapts based on the input by dynamically varying its window size. In particular, at each step, our algorithm estimates the current accuracies of the weak supervision sources by identifying a window of past observations that guarantees a near-optimal minimization of the trade-off between the error due to the variance of the estimation and the error due to the drift. Experiments on synthetic and real-world labelers show that our approach adapts to the drift.

We present a novel technique for computing a center-based approximation of a drifting distribution. Given $k \geq 1$ and a stream of data, whose distribution is changing over time, the goal is to compute, at each step, the best $k$ centers representation of the current distribution, despite possibly having only a single sample from the most recent distribution. In data mining, this is traditionally attempted through the sliding-window mechanism, where the analysis is performed on the most recent fixed-size segment of the data. The problems with this approach are twofold: (1) setting the correct window size is challenging; and (2) a fixed window size cannot effectively track changes in the distribution happening at variable speed. In this paper, we propose a new methodology that dynamically adjusts the window size based on the recent drift of the data. The challenge is that it is not possible to explicitly estimate the drift, as we may have only a single data point from each distribution. Our main contribution lies in providing a rigorous mathematical analysis, establishing both an upper bound via a dynamic window size algorithm, and a lower bound that shows the tightness of our approach.

We present a new adaptive algorithm for learning discrete distributions under distribution drift. In this setting, we observe a sequence of independent samples from a discrete distribution that is changing over time, and the goal is to estimate the current distribution. Since we have access to only a single sample for each time step, a good estimation requires a careful choice of the number of past samples to use. To use more samples, we must resort to samples further in the past, and we incur a drift error due to the bias introduced by the change in distribution. On the other hand, if we use a small number of past samples, we incur a large statistical error as the estimation has a high variance. We present a novel adaptive algorithm that can solve this trade-off without any prior knowledge of the drift. Unlike previous adaptive results, our algorithm characterizes the statistical error using data-dependent bounds. This technicality enables us to overcome the limitations of the previous work that require a fixed finite support whose size is known in advance and that cannot change over time. Additionally, we can obtain tighter bounds depending on the complexity of the drifting distribution, and also consider distributions with infinite support.

Clustering is a pivotal primitive for unsupervised learning and data analysis. A popular variant is the (k,ℓ)-clustering problem, where, given a pointset P from a metric space, one must determine a subset S of k centers minimizing the sum of the ℓ-th powers of the distances of points in P from their closest centers. This formulation covers the well-studied k-median (ℓ=1) and k-means (ℓ=2) clustering problems. A more general variant, introduced to deal with noisy pointsets, features a further parameter z and allows up to z points of P (outliers) to be disregarded when computing the sum. We present a distributed coreset-based 3-round approximation algorithm for the (k,ℓ)-clustering problem with z outliers, using MapReduce as a computational model. An important feature of our algorithm is that it obliviously adapts to the intrinsic complexity of the dataset, captured by its doubling dimension D. Remarkably, for D=O(1), our algorithm requires sublinear local memory per reducer, and yields a solution whose approximation ratio is an additive term O(γ) away from the one achievable by the best known sequential (possibly bicriteria) algorithm, where γ can be made arbitrarily small. To the best of our knowledge, no previous distributed approaches were able to attain similar quality-performance tradeoffs for metrics with constant doubling dimension.

We study nonparametric density estimation in non-stationary drift settings. Given a sequence of independent samples taken from a distribution that gradually changes in time, the goal is to compute the best estimate for the current distribution. We prove tight minimax risk bounds for both discrete and continuous smooth densities, where the minimum is over all possible estimates and the maximum is over all possible distributions that satisfy the drift constraints. Our technique handles a broad class of drift models and generalizes previous results on agnostic learning under drift.

We develop and analyze a general technique for learning with an unknown distribution drift. Given a sequence of independent observations from the last $T$ steps of a drifting distribution, our algorithm agnostically learns a family of functions with respect to the current distribution at time $T$. Unlike previous work, our technique does not require prior knowledge about the magnitude of the drift. Instead, the algorithm adapts to the sample data. Without explicitly estimating the drift, the algorithm learns a family of functions with almost the same error as a learning algorithm that knows the magnitude of the drift in advance. Furthermore, since our algorithm adapts to the data, it can guarantee a better learning error than an algorithm that relies on loose bounds on the drift. We demonstrate the application of our technique in two fundamental learning scenarios: binary classification and linear regression.

We develop a rigorous mathematical analysis of zero-shot learning with attributes. In this setting, the goal is to label novel classes with no training data, only detectors for attributes and a description of how those attributes are correlated with the target classes, called the class-attribute matrix. We develop the first non-trivial lower bound on the worst-case error of the best map from attributes to classes for this setting, even with perfect attribute detectors. The lower bound characterizes the theoretical intrinsic difficulty of the zero-shot problem based on the available information---the class-attribute matrix---and the bound is practically computable from it. Our lower bound is tight, as we show that we can always find a randomized map from attributes to classes whose expected error is upper bounded by the value of the lower bound. We show that our analysis can be predictive of how standard zero-shot methods behave in practice, including which classes will likely be confused with others.

We develop a novel method that provides theoretical guarantees for learning from weak labelers without the (mostly unrealistic) assumption that the errors of the weak labelers are independent or come from a particular family of distributions. We show a rigorous technique for efficiently selecting small subsets of the labelers so that a majority vote from such subsets has a provably low error rate. We explore several extensions of this method and provide experimental results over a range of labeled data set sizes on 45 image classification tasks. Our performance-guaranteed methods consistently match the best performing alternative, which varies based on problem difficulty. On tasks with accurate weak labelers, our methods are on average 3 percentage points more accurate than the state-of-the-art adversarial method. On tasks with inaccurate weak labelers, our methods are on average 15 percentage points more accurate than the semi-supervised Dawid-Skene model (which assumes independence).

We develop a rigorous approach for using a set of arbitrarily correlated weak supervision sources in order to solve a multiclass classification task when only a very small set of labeled data is available. Our learning algorithm provably converges to a model that has minimum empirical risk with respect to an adversarial choice over feasible labelings for a set of unlabeled data, where the feasibility of a labeling is computed through constraints defined by rigorously estimated statistics of the weak supervision sources. We show theoretical guarantees for this approach that depend on the information provided by the weak supervision sources. Notably, this method does not require the weak supervision sources to have the same labeling space as the multiclass classification task. We demonstrate the effectiveness of our approach with experiments on various image classification tasks.

Center-based clustering is a fundamental primitive for data analysis and becomes very challenging for large datasets. In this paper, we focus on the popular k-median and k-means variants which, given a set P of points from a metric space and a parameter k<|P|, require to identify a set S of k centers minimizing, respectively, the sum of the distances and of the squared distances of all points in P from their closest centers. Our specific focus is on general metric spaces, for which it is reasonable to require that the centers belong to the input set (i.e., S subseteq P). We present coreset-based 3-round distributed approximation algorithms for the above problems using the MapReduce computational model. The algorithms are rather simple and obliviously adapt to the intrinsic complexity of the dataset, captured by the doubling dimension D of the metric space. Remarkably, the algorithms attain approximation ratios that can be made arbitrarily close to those achievable by the best known polynomial-time sequential approximations, and they are very space efficient for small D, requiring local memory sizes substantially sublinear in the input size. To the best of our knowledge, no previous distributed approaches were able to attain similar quality-performance guarantees in general metric spaces.