We present an approach for building an active agent that learns to segment its visual observations into individual objects by interacting with its environment in a completely self-supervised manner. The agent uses its current segmentation model to infer pixels that constitute objects and refines the segmentation model by interacting with these pixels. The model learned from over 50K interactions generalizes to novel objects and backgrounds. To deal with noisy training signal for segmenting objects obtained by self-supervised interactions, we propose robust set loss. A dataset of robot's interactions along-with a few human labeled examples is provided as a benchmark for future research. We test the utility of the learned segmentation model by providing results on a downstream vision-based control task of rearranging multiple objects into target configurations from visual inputs alone. Videos, code, and robotic interaction dataset are available at https://pathak22.

We propose Quootstrap, a method for extracting quotations, as well as the names of the speakers who uttered them, from large news corpora. Whereas prior work has addressed this problem primarily with supervised machine learning, our approach follows a fully unsupervised bootstrapping paradigm. It leverages the redundancy present in large news corpora, more precisely, the fact that the same quotation often appears across multiple news articles in slightly different contexts. Starting from a few seed patterns, such as ["Q", said S.], our method extracts a set of quotation-speaker pairs (Q, S), which are in turn used for discovering new patterns expressing the same quotations; the process is then repeated with the larger pattern set. Our algorithm is highly scalable, which we demonstrate by running it on the large ICWSM 2011 Spinn3r corpus. Validating our results against a crowdsourced ground truth, we obtain 90% precision at 40% recall using a single seed pattern, with significantly higher recall values for more frequently reported (and thus likely more interesting) quotations. Finally, we showcase the usefulness of our algorithm's output for computational social science by analyzing the sentiment expressed in our extracted quotations.

Computable Stein discrepancies have been deployed for a variety of applications, including sampler selection in posterior inference, approximate Bayesian inference, and goodness-of-fit testing. Existing convergence-determining Stein discrepancies admit strong theoretical guarantees but suffer from a computational cost that grows quadratically in the sample size. While linear-time Stein discrepancies have been proposed for goodness-of-fit testing, they exhibit avoidable degradations in testing power---even when power is explicitly optimized. To address these shortcomings, we introduce feature Stein discrepancies ($\Phi$SDs), a new family of quality measures that can be cheaply approximated using importance sampling. We show how to construct $\Phi$SDs that provably determine the convergence of a sample to its target and develop high-accuracy approximations---random $\Phi$SDs (R$\Phi$SDs)---which are computable in near-linear time. In our experiments with sampler selection for approximate posterior inference and goodness-of-fit testing, R$\Phi$SDs typically perform as well or better than quadratic-time KSDs while being orders of magnitude faster to compute.

Autism spectrum disorder (ASD) is a developmental disorder which is currently only diagnosed through behavioral testing. Impaired folate‐dependent one carbon metabolism (FOCM) and transsulfuration (TS) pathways have been implicated in ASD, and recently a study involving multivariate analysis based upon Fisher Discriminant Analysis returned very promising results for predicting an ASD diagnosis. This article takes another step toward the goal of developing a biochemical diagnostic for ASD by comparing five classification algorithms on existing data of FOCM/TS metabolites, and also validating the classification results with new data from an ASD cohort. The comparison results indicate a high sensitivity and specificity for the original data set and up to a 88% correct classification of the ASD cohort at an expected 5% misclassification rate for typically‐developing controls. These results form the foundation for the development of a biochemical test for ASD which promises to aid diagnosis of ASD and provide biochemical understanding of the disease, applicable to at least a subset of the ASD population.

Event detection is a critical feature in data-driven systems as it assists with the identification of nominal and anomalous behavior. Event detection is increasingly relevant in robotics as robots operate with greater autonomy in increasingly unstructured environments. In this work, we present an accurate, robust, fast, and versatile measure for skill and anomaly identification. A theoretical proof establishes the link between the derivative of the log-likelihood of the HMM filtered belief state and the latest emission probabilities. The key insight is the inverse relationship in which gradient analysis is used for skill and anomaly identification. Our measure showed better performance across all metrics than related state-of-the art works. The result is broadly applicable to domains that use HMMs for event detection.

K-fold cross validation (CV) is a popular method for estimating the true performance of machine learning models, allowing model selection and parameter tuning. However, the very process of CV requires random partitioning of the data and so our performance estimates are in fact stochastic, with variability that can be substantial for natural language processing tasks. We demonstrate that these unstable estimates cannot be relied upon for effective parameter tuning. The resulting tuned parameters are highly sensitive to how our data is partitioned, meaning that we often select sub-optimal parameter choices and have serious reproducibility issues. Instead, we propose to use the less variable J-K-fold CV, in which J independent K-fold cross validations are used to assess performance. Our main contributions are extending J-K-fold CV from performance estimation to parameter tuning and investigating how to choose J and K. We argue that variability is more important than bias for effective tuning and so advocate lower choices of K than are typically seen in the NLP literature, instead use the saved computation to increase J. To demonstrate the generality of our recommendations we investigate a wide range of case-studies: sentiment classification (both general and target-specific), part-of-speech tagging and document classification.

Graph embeddings have become a key and widely used technique within the field of graph mining, proving to be successful across a broad range of domains including social, citation, transportation and biological. Graph embedding techniques aim to automatically create a low-dimensional representation of a given graph, which captures key structural elements in the resulting embedding space. However, to date, there has been little work exploring exactly which topological structures are being learned in the embeddings process. In this paper, we investigate if graph embeddings are approximating something analogous with traditional vertex level graph features. If such a relationship can be found, it could be used to provide a theoretical insight into how graph embedding approaches function. We perform this investigation by predicting known topological features, using supervised and unsupervised methods, directly from the embedding space. If a mapping between the embeddings and topological features can be found, then we argue that the structural information encapsulated by the features is represented in the embedding space. To explore this, we present extensive experimental evaluation from five state-of-the-art unsupervised graph embedding techniques, across a range of empirical graph datasets, measuring a selection of topological features. We demonstrate that several topological features are indeed being approximated by the embedding space, allowing key insight into how graph embeddings create good representations.

Traditional process monitoring techniques provide dashboards and reports showing the recent performance of a business process in terms of key performance indicators such as mean execution time, resource utilization or error rate with respect to a given notion of error. Predictive (business) process monitoring techniques go beyond traditional ones by making predictions about the future state of the executions of a business process (herein called cases). For example, a predictive monitoring technique may seek to predict the remaining execution time of each ongoing case of a process [29], the next activity that will be executed in each case [11], or the final outcome of a case, with respect to a possible set of business outcomes [23-25]. For instance, in an order-to-cash process (a process going from the receipt of a purchase order to the receipt of payment of the corresponding invoice), the possible outcomes of a case may be that the purchase order is closed satisfactorily (i.e., the customer accepted the products and paid) or unsatisfactorily (e.g., the order was canceled or withdrawn). Another set of possible outcomes is that the products were delivered on time (with respect to a maximum acceptable delivery time), or delivered late. Recent years have seen the emergence of a rich field of proposed methods for predictive process monitoring in general, and predictive monitoring of (categorical) case outcomes in particular - herein called outcome-oriented predictive process monitoring. Unfortunately, there is no unified approach to evaluate these methods. Indeed, different authors have used different datasets, experimental settings, evaluation measures and baselines.

To connect with potential customers, our marketers and sellers at Microsoft depend on good-quality leads. But sometimes people fill out online forms with fake names, gibberish, or even profanity. We distinguish fake company names from legitimate names in our data using the programming language R, the Naive Bayes classifier algorithm, Microsoft Machine Learning Server, and a data quality service that we built. This solution helps us weed out fake names and prioritize good leads for our sales and marketing teams.

We present a framework for testing independence between two random vectors that is scalable to massive data. Taking a "divide-and-conquer" approach, we break down the nonparametric multivariate test of independence into simple univariate independence tests on a collection of $2\times 2$ contingency tables, constructed by sequentially discretizing the original sample space at a cascade of scales from coarse to fine. This transforms a complex nonparametric testing problem---that traditionally requires quadratic computational complexity with respect to the sample size---into a multiple testing problem that can be addressed with a computational complexity that scales almost linearly with the sample size. We further consider the scenario when the dimensionality of the two random vectors also grows large, in which case the curse of dimensionality arises in the proposed framework through an explosion in the number of univariate tests to be completed. To overcome this difficulty, we propose a data-adaptive version of our method that completes a fraction of the univariate tests, judged to be more likely to contain evidence for dependency based on exploiting the spatial characteristics of the dependency structure in the data. We provide an inference recipe based on multiple testing adjustment that guarantees the inferential validity in terms of properly controlling the family-wise error rate. We demonstrate the tremendous computational advantage of the algorithm in comparison to existing approaches while achieving desirable statistical power through an extensive simulation study. In addition, we illustrate how our method can be used for learning the nature of the underlying dependency in addition to hypothesis testing. We demonstrate the use of our method through analyzing a data set from flow cytometry.