We propose a new active learning algorithm for parametric linear regression with random design. We provide finite sample convergence guarantees for general distributions in the misspecified model. This is the first active learner for this setting that provably can improve over passive learning. Unlike other learning settings (such as classification), in regression the passive learning rate of $O(1/\epsilon)$ cannot in general be improved upon. Nonetheless, the so-called `constant' in the rate of convergence, which is characterized by a distribution-dependent risk, can be improved in many cases. For a given distribution, achieving the optimal risk requires prior knowledge of the distribution. Following the stratification technique advocated in Monte-Carlo function integration, our active learner approaches the optimal risk using piecewise constant approximations.
We explore how to improve machine translation systems by adding more translation data in situations where we already have substantial resources. The main challenge is how to buck the trend of diminishing returns that is commonly encountered. We present an active learning-style data solicitation algorithm to meet this challenge. We test it, gathering annotations via Amazon Mechanical Turk, and find that we get an order of magnitude increase in performance rates of improvement.
We propose a novel sampling framework for inference in probabilistic models: an active learning approach that converges more quickly (in wall-clock time) than Markov chain Monte Carlo (MCMC) benchmarks. The central challenge in probabilistic inference is numerical integration, to average over ensembles of models or unknown (hyper-)parameters (for example to compute the marginal likelihood or a partition function). MCMC has provided approaches to numerical integration that deliver state-of-the-art inference, but can suffer from sample inefficiency and poor convergence diagnostics. Bayesian quadrature techniques offer a model-based solution to such problems, but their uptake has been hindered by prohibitive computation costs. We introduce a warped model for probabilistic integrands (likelihoods) that are known to be non-negative, permitting a cheap active learning scheme to optimally select sample locations. Our algorithm is demonstrated to offer faster convergence (in seconds) relative to simple Monte Carlo and annealed importance sampling on both synthetic and real-world examples.
We describe a framework for designing efficient active learning algorithms that are tolerant to random classification noise and are differentially-private. The framework is based on active learning algorithms that are statistical in the sense that they rely on estimates of expectations of functions of filtered random examples. It builds on the powerful statistical query framework of Kearns (1993). We show that any efficient active statistical learning algorithm can be automatically converted to an efficient active learning algorithm which is tolerant to random classification noise as well as other forms of "uncorrelated" noise. The complexity of the resulting algorithms has information-theoretically optimal quadratic dependence on $1/(1-2\eta)$, where $\eta$ is the noise rate. We show that commonly studied concept classes including thresholds, rectangles, and linear separators can be efficiently actively learned in our framework. These results combined with our generic conversion lead to the first computationally-efficient algorithms for actively learning some of these concept classes in the presence of random classification noise that provide exponential improvement in the dependence on the error $\epsilon$ over their passive counterparts. In addition, we show that our algorithms can be automatically converted to efficient active differentially-private algorithms. This leads to the first differentially-private active learning algorithms with exponential label savings over the passive case.
Crowd-sourcing has become a popular means of acquiring labeled data for a wide variety of tasks where humans are more accurate than computers, e.g., labeling images, matching objects, or analyzing sentiment. However, relying solely on the crowd is often impractical even for data sets with thousands of items, due to time and cost constraints of acquiring human input (which cost pennies and minutes per label). In this paper, we propose algorithms for integrating machine learning into crowd-sourced databases, with the goal of allowing crowd-sourcing applications to scale, i.e., to handle larger datasets at lower costs. The key observation is that, in many of the above tasks, humans and machine learning algorithms can be complementary, as humans are often more accurate but slow and expensive, while algorithms are usually less accurate, but faster and cheaper. Based on this observation, we present two new active learning algorithms to combine humans and algorithms together in a crowd-sourced database. Our algorithms are based on the theory of non-parametric bootstrap, which makes our results applicable to a broad class of machine learning models. Our results, on three real-life datasets collected with Amazon's Mechanical Turk, and on 15 well-known UCI data sets, show that our methods on average ask humans to label one to two orders of magnitude fewer items to achieve the same accuracy as a baseline that labels random images, and two to eight times fewer questions than previous active learning schemes.
A central question for active learning (AL) is: "what is the optimal selection?" Defining optimality by classifier loss produces a new characterisation of optimal AL behaviour, by treating expected loss reduction as a statistical target for estimation. This target forms the basis of model retraining improvement (MRI), a novel approach providing a statistical estimation framework for AL. This framework is constructed to address the central question of AL optimality, and to motivate the design of estimation algorithms. MRI allows the exploration of optimal AL behaviour, and the examination of AL heuristics, showing precisely how they make sub-optimal selections. The abstract formulation of MRI is used to provide a new guarantee for AL, that an unbiased MRI estimator should outperform random selection. This MRI framework reveals intricate estimation issues that in turn motivate the construction of new statistical AL algorithms. One new algorithm in particular performs strongly in a large-scale experimental study, compared to standard AL methods. This competitive performance suggests that practical efforts to minimise estimation bias may be important for AL applications.
From a machine learning perspective, the human ability localize sounds can be modeled as a non-parametric and non-linear regression problem between binaural spectral features of sound received at the ears (input) and their sound-source directions (output). The input features can be summarized in terms of the individual's head-related transfer functions (HRTFs) which measure the spectral response between the listener's eardrum and an external point in $3$D. Based on these viewpoints, two related problems are considered: how can one achieve an optimal sampling of measurements for training sound-source localization (SSL) models, and how can SSL models be used to infer the subject's HRTFs in listening tests. First, we develop a class of binaural SSL models based on Gaussian process regression and solve a \emph{forward selection} problem that finds a subset of input-output samples that best generalize to all SSL directions. Second, we use an \emph{active-learning} approach that updates an online SSL model for inferring the subject's SSL errors via headphones and a graphical user interface. Experiments show that only a small fraction of HRTFs are required for $5^{\circ}$ localization accuracy and that the learned HRTFs are localized closer to their intended directions than non-individualized HRTFs.
We study the learnability of linear separators in $\Re^d$ in the presence of bounded (a.k.a Massart) noise. This is a realistic generalization of the random classification noise model, where the adversary can flip each example $x$ with probability $\eta(x) \leq \eta$. We provide the first polynomial time algorithm that can learn linear separators to arbitrarily small excess error in this noise model under the uniform distribution over the unit ball in $\Re^d$, for some constant value of $\eta$. While widely studied in the statistical learning theory community in the context of getting faster convergence rates, computationally efficient algorithms in this model had remained elusive. Our work provides the first evidence that one can indeed design algorithms achieving arbitrarily small excess error in polynomial time under this realistic noise model and thus opens up a new and exciting line of research. We additionally provide lower bounds showing that popular algorithms such as hinge loss minimization and averaging cannot lead to arbitrarily small excess error under Massart noise, even under the uniform distribution. Our work instead, makes use of a margin based technique developed in the context of active learning. As a result, our algorithm is also an active learning algorithm with label complexity that is only a logarithmic the desired excess error $\epsilon$.
In chemometrics, data from infrared or near-infrared (NIR) spectroscopy are often used to identify a compound or to analyze the composition of amaterial. This involves the calibration of models that predict the concentration ofmaterial constituents from the measured NIR spectrum. An interesting aspect of multivariate calibration is to achieve a particular accuracy level with a minimum number of training samples, as this reduces the number of laboratory tests and thus the cost of model building. In these chemometric models, the input refers to a proper representation of the spectra and the output to the concentrations of the sample constituents. The search for a most informative new calibration sample thus has to be performed in the output space of the model, rather than in the input space as in conventionalmodeling problems. In this paper, we propose to solve the corresponding inversion problem by utilizing the disagreements of an ensemble of neural networks to represent the prediction error in the unexplored component space. The next calibration sample is then chosen at a composition where the individual models of the ensemble disagree most. The results obtained for a realistic chemometric calibration example show that the proposed active learning can achieve a given calibration accuracy with less training samples than random sampling.
Active learning is a type of sequential design for supervised machine learning, in which the learning algorithm sequentially requests the labels of selected instances from a large pool of unlabeled data points. The objective is to produce a classifier of relatively low risk, as measured under the 0-1 loss, ideally using fewer label requests than the number of random labeled data points sufficient to achieve the same. This work investigates the potential uses of surrogate loss functions in the context of active learning. Specifically, it presents an active learning algorithm based on an arbitrary classification-calibrated surrogate loss function, along with an analysis of the number of label requests sufficient for the classifier returned by the algorithm to achieve a given risk under the 0-1 loss. Interestingly, these results cannot be obtained by simply optimizing the surrogate risk via active learning to an extent sufficient to provide a guarantee on the 0-1 loss, as is common practice in the analysis of surrogate losses for passive learning. Some of the results have additional implications for the use of surrogate losses in passive learning.
We consider the problem of learning convex aggregation of models, that is as good as the best convex aggregation, for the binary classification problem. Working in the stream based active learning setting, where the active learner has to make a decision on-the-fly, if it wants to query for the label of the point currently seen in the stream, we propose a stochastic-mirror descent algorithm, called SMD-AMA, with entropy regularization. We establish an excess risk bounds for the loss of the convex aggregate returned by SMD-AMA to be of the order of $O\left(\sqrt{\frac{\log(M)}{{T^{1-\mu}}}}\right)$, where $\mu\in [0,1)$ is an algorithm dependent parameter, that trades-off the number of labels queried, and excess risk.
Active learning has shown to reduce the number of experiments needed to obtain high-confidence drug-target predictions. However, in order to actually save experiments using active learning, it is crucial to have a method to evaluate the quality of the current prediction and decide when to stop the experimentation process. Only by applying reliable stoping criteria to active learning, time and costs in the experimental process can be actually saved. We compute active learning traces on simulated drug-target matrices in order to learn a regression model for the accuracy of the active learner. By analyzing the performance of the regression model on simulated data, we design stopping criteria for previously unseen experimental matrices. We demonstrate on four previously characterized drug effect data sets that applying the stopping criteria can result in upto 40% savings of the total experiments for highly accurate predictions.
In order to assist security analysts in obtaining information pertaining to their network, such as novel vulnerabilities, exploits, or patches, information retrieval methods tailored to the security domain are needed. As labeled text data is scarce and expensive, we follow developments in semi-supervised Natural Language Processing and implement a bootstrapping algorithm for extracting security entities and their relationships from text. The algorithm requires little input data, specifically, a few relations or patterns (heuristics for identifying relations), and incorporates an active learning component which queries the user on the most important decisions to prevent drifting from the desired relations. Preliminary testing on a small corpus shows promising results, obtaining precision of .82.
Conventional multimedia annotation/retrieval systems such as Normalized Continuous Relevance Model (NormCRM) [16] require a fully labeled training data for a good performance. Active Learning, by determining an order for labeling the training data, allows for a good performance even before the training data is fully annotated. In this work we propose an active learning algorithm, which combines a novel measure of sample uncertainty with a novel clustering-based approach for determining sample density and diversity and integrate it with NormCRM. The clusters are also iteratively refined to ensure both feature and label-level agreement among samples. We show that our approach outperforms multiple baselines both on a recent, open character animation dataset and on the popular TRECVID corpus at both the tasks of annotation and text-based retrieval of videos.
We enhance the autonomy of the continuous active learning method shown by Cormack and Grossman (SIGIR 2014) to be effective for technology-assisted review, in which documents from a collection are retrieved and reviewed, using relevance feedback, until substantially all of the relevant documents have been reviewed. Autonomy is enhanced through the elimination of topic-specific and dataset-specific tuning parameters, so that the sole input required by the user is, at the outset, a short query, topic description, or single relevant document; and, throughout the review, ongoing relevance assessments of the retrieved documents. We show that our enhancements consistently yield superior results to Cormack and Grossman's version of continuous active learning, and other methods, not only on average, but on the vast majority of topics from four separate sets of tasks: the legal datasets examined by Cormack and Grossman, the Reuters RCV1-v2 subject categories, the TREC 6 AdHoc task, and the construction of the TREC 2002 filtering test collection.
Within the natural language processing (NLP) community, active learning has been widely investigated and applied in order to alleviate the annotation bottleneck faced by developers of new NLP systems and technologies. This paper presents the first theoretical analysis of stopping active learning based on stabilizing predictions (SP). The analysis has revealed three elements that are central to the success of the SP method: (1) bounds on Cohen's Kappa agreement between successively trained models impose bounds on differences in F-measure performance of the models; (2) since the stop set does not have to be labeled, it can be made large in practice, helping to guarantee that the results transfer to previously unseen streams of examples at test/application time; and (3) good (low variance) sample estimates of Kappa between successive models can be obtained. Proofs of relationships between the level of Kappa agreement and the difference in performance between consecutive models are presented. Specifically, if the Kappa agreement between two models exceeds a threshold T (where $T>0$), then the difference in F-measure performance between those models is bounded above by $\frac{4(1-T)}{T}$ in all cases. If precision of the positive conjunction of the models is assumed to be $p$, then the bound can be tightened to $\frac{4(1-T)}{(p+1)T}$.
To train good supervised and semi-supervised object classifiers, it is critical that we not waste the time of the human experts who are providing the training labels. Existing active learning strategies can have uneven performance, being efficient on some datasets but wasteful on others, or inconsistent just between runs on the same dataset. We propose perplexity based graph construction and a new hierarchical subquery evaluation algorithm to combat this variability, and to release the potential of Expected Error Reduction. Under some specific circumstances, Expected Error Reduction has been one of the strongest-performing informativeness criteria for active learning. Until now, it has also been prohibitively costly to compute for sizeable datasets. We demonstrate our highly practical algorithm, comparing it to other active learning measures on classification datasets that vary in sparsity, dimensionality, and size. Our algorithm is consistent over multiple runs and achieves high accuracy, while querying the human expert for labels at a frequency that matches their desired time budget.
In active learning, the user sequentially chooses values for feature $X$ and an oracle returns the corresponding label $Y$. In this paper, we consider the effect of feature noise in active learning, which could arise either because $X$ itself is being measured, or it is corrupted in transmission to the oracle, or the oracle returns the label of a noisy version of the query point. In statistics, feature noise is known as "errors in variables" and has been studied extensively in non-active settings. However, the effect of feature noise in active learning has not been studied before. We consider the well-known Berkson errors-in-variables model with additive uniform noise of width $\sigma$. Our simple but revealing setting is that of one-dimensional binary classification setting where the goal is to learn a threshold (point where the probability of a $+$ label crosses half). We deal with regression functions that are antisymmetric in a region of size $\sigma$ around the threshold and also satisfy Tsybakov's margin condition around the threshold. We prove minimax lower and upper bounds which demonstrate that when $\sigma$ is smaller than the minimiax active/passive noiseless error derived in \cite{CN07}, then noise has no effect on the rates and one achieves the same noiseless rates. For larger $\sigma$, the \textit{unflattening} of the regression function on convolution with uniform noise, along with its local antisymmetry around the threshold, together yield a behaviour where noise \textit{appears} to be beneficial. Our key result is that active learning can buy significant improvement over a passive strategy even in the presence of feature noise.
Interesting theoretical associations have been established by recent papers between the fields of active learning and stochastic convex optimization due to the common role of feedback in sequential querying mechanisms. In this paper, we continue this thread in two parts by exploiting these relations for the first time to yield novel algorithms in both fields, further motivating the study of their intersection. First, inspired by a recent optimization algorithm that was adaptive to unknown uniform convexity parameters, we present a new active learning algorithm for one-dimensional thresholds that can yield minimax rates by adapting to unknown noise parameters. Next, we show that one can perform $d$-dimensional stochastic minimization of smooth uniformly convex functions when only granted oracle access to noisy gradient signs along any coordinate instead of real-valued gradients, by using a simple randomized coordinate descent procedure where each line search can be solved by $1$-dimensional active learning, provably achieving the same error convergence rate as having the entire real-valued gradient. Combining these two parts yields an algorithm that solves stochastic convex optimization of uniformly convex and smooth functions using only noisy gradient signs by repeatedly performing active learning, achieves optimal rates and is adaptive to all unknown convexity and smoothness parameters.
We study the problem of learning in the presence of a drifting target concept. Specifically, we provide bounds on the error rate at a given time, given a learner with access to a history of independent samples labeled according to a target concept that can change on each round. One of our main contributions is a refinement of the best previous results for polynomial-time algorithms for the space of linear separators under a uniform distribution. We also provide general results for an algorithm capable of adapting to a variable rate of drift of the target concept. Some of the results also describe an active learning variant of this setting, and provide bounds on the number of queries for the labels of points in the sequence sufficient to obtain the stated bounds on the error rates.