Instrumental variable regression (IVR) is a statistical technique utilized to recover unbiased estimators when there are errors in the independent variables. Estimator bias in learned time series models can yield poor performance in applications such as long-term prediction and filtering where the recursive use of the model results in the accumulation of propagated error. However, prior work addressed the IVR objective in the batch setting, where it is necessary to store the entire dataset in memory — an infeasible requirementin large dataset scenarios. In this work, we develop Online Instrumental Variable Regression (OIVR), an algorithm that is capable of updating the learned estimator with streaming data. We show that the online adaptation of IVR enjoys a no-regret performance guarantee with respect the original batchsetting by taking advantage of any no-regret online learning algorithm inside OIVR for the underlying update steps. We experimentally demonstrate the efficacy of our algorithm in combination with popular no-regret onlinealgorithms for the task of learning predictive dynamical system models and on a prototypical econometrics instrumental variable regression problem.

Kaczmarz algorithm is an efficient iterative algorithm to solve overdetermined consistent system of linear equations. During each updating step, Kaczmarz chooses a hyperplane based on an individual equation and projects the current estimate for the exact solution onto that space to get a new estimate.Many vairants of Kaczmarz algorithms are proposed on how to choose better hyperplanes.Using the property of randomly sampled data in high-dimensional space,we propose an accelerated algorithm based on clustering information to improve block Kaczmarz and Kaczmarz via Johnson-Lindenstrauss lemma. Additionally, we theoretically demonstrate convergence improvement on block Kaczmarz algorithm.

Many interesting machine learning problems are best posed by considering instances that are distributions, or sample sets drawn from distributions. Most previous work devoted to machine learning tasks with distributional inputs has done so through pairwise kernel evaluations between pdfs (or sample sets). While such an approach is fine for smaller datasets, the computation of an N × N Gram matrix is prohibitive in large datasets. Recent scalable estimators that work over pdfs have done so only with kernels that use Euclidean metrics, like the L 2 distance. However, there are a myriad of other useful metrics available, such as total variation, Hellinger distance, and the Jensen-Shannon divergence. This work develops the first random features for pdfs whose dot product approximates kernels using these non-Euclidean metrics. These random features allow estimators to scale to large datasets by working in a primal space, without computing large Gram matrices. We provide an analysis of the approximation error in using our proposed random features, and show empirically the quality of our approximation both in estimating a Gram matrix and in solving learning tasks in real-world and synthetic data.

Feature extraction is an important task in machine learning. In this paper, we present a simple and efficient method, named max-margin data shifting (MMDS), to process the data before feature extraction. By relying on a large-margin classifier, MMDS is helpful to enhance the discriminative ability of subsequent feature extractors. The kernel trick can be applied to extract nonlinear features from input data. We further analyze in detail the example of principal component analysis (PCA). The empirical results on multiple linear and nonlinear models demonstrate that MMDS can efficiently improve the performance of unsupervised extractors.

Distant labeling for information extraction (IE) suffers from noisy training data. We describe a way of reducing the noise associated with distant IE by identifying coupling constraints between potential instance labels. As one example of coupling,items in a list are likely to have the same label.A second example of coupling comes from analysis of document structure: in some corpora,sections can be identified such that items in the same section are likely to have the same label. Such sections do not exist in all corpora, but we show that augmenting a large corpus with coupling constraints from even a small, well-structured corpus can improve performance substantially, doubling F1 on one task.

Survival prediction is crucial to healthcare research, but is confined primarily to specific types of data involving only the present measurements. This paper considers the more general class of healthcare data found in practice, which includes a wealth of intermittently varying historical measurements in addition to the present measurements. Making survival predictions on such data bristles with challenges to the existing prediction models. For this reason, we propose a new semi-proportional hazards model using locally time-varying coefficients, and a novel complete-data model learning criterion for coefficient optimization. Experiments on the healthcare data demonstrate the effectiveness and generalizability of our model and its promise in practical applications.

ICU mortality risk prediction may help clinicians take effective interventions to improve patient outcome. Existing machine learning approaches often face challenges in integrating a comprehensive panel of physiologic variables and presenting to clinicians interpretable models. We aim to improve both accuracy and interpretability of prediction models by introducing Subgraph Augmented Non-negative Matrix Factorization (SANMF) on ICU physiologic time series. SANMF converts time series into a graph representation and applies frequent subgraph mining to automatically extract temporal trends. We then apply non-negative matrix factorization to group trends in a way that approximates patient pathophysiologic states. Trend groups are then used as features in training a logistic regression model for mortality risk prediction, and are also ranked according to their contribution to mortality risk. We evaluated SANMF against four empirical models on the task of predicting mortality or survival 30 days after discharge from ICU using the observed physiologic measurements between 12 and 24 hours after admission. SANMF outperforms all comparison models, and in particular, demonstrates an improvement in AUC (0.848 vs. 0.827, p<0.002) compared to a state-of-the-art machine learning method that uses manual feature engineering. Feature analysis was performed to illuminate insights and benefits of subgraph groups in mortality risk prediction.

We consider an online learning problem (classification or prediction) involving disparate sources of sequentially arriving data, whereby a user over time learns the best set of data sources to use in constructing the classifier by exploiting their similarity. We first show that, when (1) the similarity information among data sources is known, and (2) data from different sources can be acquired without cost, then a judicious selection of data from different sources can effectively enlarge the training sample size compared to using a single data source, thereby improving the rate and performance of learning; this is achieved by bounding the classification error of the resulting classifier. We then relax assumption (1) and characterize the loss in learning performance when the similarity information must also be acquired through repeated sampling. We further relax both (1) and (2) and present a cost-efficient algorithm that identifies a best crowd from a potentially large set of data sources in terms of both classifier performance and data acquisition cost. This problem has various applications, including online prediction systems with time series data of various forms, such as financial markets, advertisement and network measurement.

Utilizing trajectories for modeling human mobility often involves extracting descriptive features for each individual, a procedure heavily based on experts' knowledge. In this work, our objective is to minimize human involvement and exploit the power of community in learning `features' for individuals from their location traces. We propose a probabilistic graphical model that learns distribution of latent concepts, named motifs, from anonymized sequences of user locations. To handle variation in user activity level, our model learns motif distributions from sequence-level location co-occurrence of all users. To handle the big variation in location popularity, our model uses an asymmetric prior, conditioned on per-sequence features. We evaluate the new representation in a link prediction task and compare our results to those of baseline approaches.

Computational agents often need to learn policies that involve many control variables, e.g., a robot needs to control several joints simultaneously. Learning a policy with a high number of parameters, however, usually requires a large number of training samples. We introduce a reinforcement learning method for sample-efficient policy search that exploits correlations between control variables. Such correlations are particularly frequent in motor skill learning tasks. The introduced method uses Variational Inference to estimate policy parameters, while at the same time uncovering a low-dimensional latent space of controls. Prior knowledge about the task and the structure of the learning agent can be provided by specifying groups of potentially correlated parameters. This information is then used to impose sparsity constraints on the mapping between the high-dimensional space of controls and a lower-dimensional latent space. In experiments with a simulated bi-manual manipulator, the new approach effectively identifies synergies between joints, performs efficient low-dimensional policy search, and outperforms state-of-the-art policy search methods.