Opinion leaders play an important role in influencing people’s beliefs, actions and behaviors. Although a number of methods have been proposed for identifying influentials using secondary sources of information, the use of primary sources, such as surveys, is still favored in many domains. In this work we present a new surveying method which combines secondary data with partial knowledge from primary sources to guide the information gathering process. We apply our proposed active surveying method to the problem of identifying key opinion leaders in the medical field, and show how we are able to accurately identify the opinion leaders while minimizing the amount of primary data required, which results in significant cost reduction in data acquisition without sacrificing its integrity.

Understanding extreme events, such as hurricanes or forest fires, is of paramount importance because of their adverse impacts on human beings. Such events often propagate in space and time. Predicting-even a few days in advance-what locations will get affected by the event tracks could benefit our society in many ways. Arguably, simulations from “first principles,” where underlying physics-based models are described by a system of equations, provide least reliable predictions for variables characterizing the dynamics of these extreme events. Data-driven model building has been recently emerging as a complementary approach that could learn the relationships between historically observed or simulated multiple, spatio-temporal ancillary variables and the dynamic behavior of extreme events of interest. While promising, the methodology for predictive learning from such complex data is still in its infancy. In this paper, we propose a dynamic networks-based methodology for in-advance prediction of the dynamic tracks of emerging extreme events. By associating a network model of the system with the known tracks, our method is capable of learning the recurrent network motifs that could be used as discriminatory signatures for the event's behavioral class. When applied to classifying the behavior of the hurricane tracks at their early formation stages in Western Africa region, our method is able to predict whether hurricane tracks will hit the land of the North Atlantic region at least 10-15 days lead lag time in advance with more than 90% accuracy using 10-fold cross-validation. To the best of our knowledge, no comparable methodology exists for solving this problem using data-driven models.

We propose a general MCMC method for Bayesian inference in logic-based probabilistic modeling. It covers a broad class of generative models including Bayesian networks and PCFGs. The idea is to generalize an MCMC method for PCFGs to the one for a Turing-complete probabilistic modeling language PRISM in the context of statistical abduction where parse trees are replaced with explanations. We describe how to estimate the marginal probability of data from MCMC samples and how to perform Bayesian Viterbi inference using an example of Naive Bayes model augmented with a hidden variable.

Continuous time series data often comprise or contain repeated motifs — patterns that have similar shape, and yet exhibit nontrivial variability. Identifying these motifs, even in the presence of variation, is an important subtask in both unsupervised knowledge discovery and constructing useful features for discriminative tasks. This paper addresses this task using a probabilistic framework that models generation of data as switching between a random walk state and states that generate motifs. A motif is generated from a continuous shape template that can undergo non-linear transformations such as temporal warping and additive noise. We propose an unsupervised algorithm that simultaneously discovers both the set of canonical shape templates and a template-specific model of variability manifested in the data. Experimental results on three real-world data sets demonstrate that our model is able to recover templates in data where repeated instances show large variability. The recovered templates provide higher classification accuracy and coverage when compared to those from alternatives such as random projection based methods and simpler generative models that do not model variability. Moreover, in analyzing physiological signals from infants in the ICU, we discover both known signatures as well as novel physiomarkers.

We present a novel approach for domain adaptation based on feature grouping and re-weighting. Our algorithm operates by creating an ensemble of multiple classifiers, where each classifier is trained on one particular feature group. Faced with the distribution change involved in domain change, different feature groups exhibit different cross-domain prediction abilities. Herein, ensemble models provide us the flexibility of tuning the weights of corresponding classifiers in order to adapt to the new domain. Our approach is supported by a solid theoretical analysis based on the expressiveness of ensemble classifiers, which allows trading-off errors across source and target domains. Moreover, experimental results on sentiment classification and spam detection show that our approach not only outperforms the baseline method, but is also superior to other state-of-the-art methods.

This paper introduces a novel multi-modular method for reinforcement learning.A multi-modular system is one that partitions the learning task among a set of experts (modules), where each expert is incapable of solving the entire task by itself.There are many advantages to splitting up large tasks in this way, but existing methods face difficulties when choosing which module(s) should contribute to the agent's actions at any particular moment.We introduce a novel selection mechanism where every module, besides calculating a set of action values, also estimates its own error for the current input.The selection mechanism combines each module's estimate of long-term reward and self-error to produce a score by which the next module is chosen.As a result, the modules can use their resources effectively and efficiently divide up the task.The system is shown to learn complex tasks even when the individual modules use only linear function approximators.

Distributed electricity producers, such as small wind farms and solar installations, pose several technical and economic challenges in Smart Grid design. One approach to addressing these challenges is through Broker Agents who buy electricity from distributed producers, and also sell electricity to consumers, via a Tariff Market--a new market mechanism where Broker Agents publish concurrent bid and ask prices. We investigate the learning of pricing strategies for an autonomous Broker Agent to profitably participate in a Tariff Market. We employ Markov Decision Processes (MDPs) and reinforcement learning. An important concern with this method is that even simple representations of the problem domain result in very large numbers of states in the MDP formulation because market prices can take nearly arbitrary real values. In this paper, we present the use of derived state space features, computed using statistics on Tariff Market prices and Broker Agent customer portfolios, to obtain a scalable state representation. We also contribute a set of pricing tactics that form building blocks in the learned Broker Agent strategy. We further present a Tariff Market simulation model based on real-world data and anticipated market dynamics. We use this model to obtain experimental results that show the learned strategy performing vastly better than a random strategy and significantly better than two other non-learning strategies.

In this paper, we present BENCH (BiclusteringdrivenENsemble of Classifiers), an algorithm toconstruct an ensemble of classifiers through concurrentfeature and data point selection guided byunsupervised knowledge obtained from biclustering.BENCH is designed for underdeterminedproblems. In our experiments, we use Bayesian BeliefNetwork (BBN) classifiers as base classifiers inthe ensemble; however, BENCH can be applied toother classification models as well. We show thatBENCH is able to increase prediction accuracy ofa single classifier and traditional ensemble of classifiersby up to 15% on three microarray datasetsusing various weighting schemes for combining individualpredictions in the ensemble.

Principal Component Analysis (PCA) is one of the most important methods to handle high-dimensional data. However, the high computa-tional complexity makes it hard to apply to the large scale data with high dimensionality, and the used ℓ2-norm makes it sensitive to outliers. A recent work proposed principal component analysis based on ℓ1-norm maximization, which is efficient and robust to outliers. In that work, a greedy strategy was applied due to the difficulty of directly solving the ℓ1-norm maximization problem, which is easy to get stuck in local solution. In this paper, we first propose an efficient optimization algorithm to solve a general ℓ1-norm maximization problem, and then propose a robust principal component analysis with non-greedy ℓ1-norm maximization. Experimental results on real world datasets show that the non-greedy method always obtains much better solution than that of the greedy method.

We propose a family of Passive-Aggressive Mahalanobis (PAM) algorithms, which are incremental (online) binary classifiers that consider the distribution of data. PAM is in fact a generalization of the Passive-Aggressive (PA) algorithms to handle data distributions that can be represented by a covariance matrix. The update equations for PAM are derived and theoretical error loss bounds computed. We benchmarked PAM against the original PA-I, PA-II, and Confidence Weighted (CW) learning. Although PAM somewhat resembles CW in its update equations, PA minimizes differences in the weights while CW minimizes differences in weight distributions. Results on 8 classification datasets, which include a real-life micro-blog sentiment classification task, show that PAM consistently outperformed its competitors, most notably CW. This shows that a simple approach like PAM is more practical in real-life classification tasks, compared to more elegant and sophisticated approaches like CW.