The Approximate Linear Programming (ALP) approach to value function approximation for MDPs is a parametric value function approximation method, in that it represents the value function as a linear combination of features which are chosen a priori. Choosing these features can be a difficult challenge in itself. One recent effort, Regularized Approximate Linear Programming (RALP), uses L1 regularization to address this issue by combining a large initial set of features with a regularization penalty that favors a smooth value function with few non-zero weights. Rather than using smoothness as a backhanded way of addressing the feature selection problem, this paper starts with smoothness and develops a non-parametric approach to ALP that is consistent with the smoothness assumption. We show that this new approach has some favorable practical and analytical properties in comparison to (R)ALP.

We present a spectrum-based, sequential software debugging approach coined Sequoia, that greedily selects tests out of a suite of tests to narrow down the set of diagnostic candidates with a minimum number of tests. Sequoia handles multiple faults, that can be intermittent, at polynomial time and space complexity, due to a novel, approximate diagnostic entropy estimation approach, which considers the subset of diagnoses that cover almost all Bayesian posterior probability mass. Synthetic experiments show that Sequoia achieves much better diagnostic uncertainty reduction compared to random test sequencing.Real programs, taken from the Software Infrastructure Repository, confirm Sequoia's better performance, with a test reduction up to 80% compared to random test sequences.

Computer Go presents a challenging problem for machine learning agents. With the number of possible board states estimated to be larger than the number of hydrogen atoms in the universe, learning effective policies or board evaluation functions is extremely difficult. In this paper we describe Cortigo, a system that efficiently and autonomously learns useful generalizations for large state-space classification problems such as Go. Cortigo uses a hierarchical generative model loosely related to the human visual cortex to recognize Go board positions well enough to suggest promising next moves. We begin by briefly describing and providing motivation for research in the computer Go domain. We describe Cortigo’s ability to learn predictive models based on large subsets of the Go board and demonstrate how using Cortigo’s learned models as additive knowledge in a state-of-the-art computer Go player (Fuego) significantly improves its playing strength.

In multiple-Instance Learning (MIL), training class labels are attached to sets of bags composed of unlabeled instances, and the goal is to deal with classification of bags. Most previous MIL algorithms, which tackle classification problems, consider each instance as a represented feature. Although the algorithms work well in some prediction problems, considering diverse features to represent an instance may provide more significant information for learning task. Moreover, since each instance may be mapped into diverse feature spaces, encountering a large number of irrelevant or redundant features is inevitable. In this paper, we propose a method to select relevant instances and concurrently consider multiple features for each instance, which is termed as MIL-MFS. MIL-MFS is based on multiple kernel learning (MKL), and it iteratively selects the fusing multiple features for classifier training. Experimental results show that the MIL-MFS combined with multiple kernel learning can significantly improve the classification performance.

Multilabel classification is a central problem in many areas of data analysis, including text and multimedia categorization, where individual data objects need to be assigned multiple labels. A key challenge in these tasks is to learn a classifier that can properly exploit label correlations without requiring exponential enumeration of label subsets during training or testing. We investigate novel loss functions for multilabel training within a large margin framework---identifying a simple alternative that yields improved generalization while still allowing efficient training. We furthermore show how covariances between the label models can be learned simultaneously with the classification model itself, in a jointly convex formulation, without compromising scalability. The resulting combination yields state of the art accuracy in multilabel webpage classification.

The aim of a modern Building Automation System (BAS) is to enhance interactive control strategies for energy efficiency and user comfort. In this context, we develop a novel control algorithm that uses a stochastic building occupancy model to improve mean energy efficiency while minimizing expected discomfort. We compare by simulation our Stochastic Model Predictive Control (SMPC) strategy to the standard heating control method to empirically demonstrate a 4.3% reduction in energy use and 38.3% reduction in expected discomfort.

This paper explores a new measure of similarity between point sets in arbitrary metric spaces. The measure is based on the spatial overlap of the “shapes” and “densities” of these point sets. It is applicable in any domain where point sets are a natural representation for data. Specifically, we show examples of its use in natural language processing, object recognition in images and point set classification. We provide a geometric interpretation of this measure and show that it is well-motivated, intuitive, parameter-free, and straightforward to use. We further demonstrate that it is computationally tractable and applicable to both supervised and unsupervised learning problems.

We focus on a class of models used for representing the dynamics between a discrete set of probabilistic events in a continuous-time setting. The proposed framework offers tractable learning and inference procedures and provides compact state representations for processes which exhibit variable delays between events. The approach is applied to a heart sound labeling task that exhibits long-range dependencies on previous events, and in which explicit modeling of the rhythm timings is justifiable by cardiological principles.

Empirical findings indicate that humans draw infer- ences about spatial arrangements by constructing and manipulating mental models which are internal representations of objects and relations in spatial working memory. Central to the Mental Model Theory (MMT), is the assumption that the human reasoning process can be divided into three phases: (i) Mental model construction, (ii) model inspection, and (iii) model validation. The MMT can be formalized with respect to a computational model, connecting the reasoning process to operations on mental model representations. In this respect a computational model has been implemented in the cognitive architecture ACT-R capable of explaining human reasoning difficulty by the number of model operations. The presented ACT-R model allows simulation of psychological findings about spatial reasoning problems from a previous study that investigated conventional behavioral data such as response times and error rates in the context of certain mental model construction principles.

Strategy mining is a new area of research about discovering strategies in decision-making. In this paper, we formulate the strategy-mining problem as a clustering problem, called the latent-strategy problem. In a latent-strategy problem, a corpus of data instances is given, each of which is represented by a set of features and a decision label. The inherent dependency of the decision label on the features is governed by a latent strategy. The objective is to find clusters, each of which contains data instances governed by the same strategy. Existing clustering algorithms are inappropriate to cluster dependency because they either assume feature independency (e.g., K-means) or only consider the co-occurrence of features without explicitly modeling the special dependency of the decision label on other features (e.g., Latent Dirichlet Allocation (LDA)). In this paper, we present a baseline unsupervised learning algorithm for dependency clustering. Our model-based clustering algorithm iterates between an assignment step and a minimization step to learn a mixture of decision tree models that represent latent strategies. Similar to the Expectation Maximization algorithm, our algorithm is grounded in the statistical learning theory. Different from other clustering algorithms, our algorithm is irrelevant-feature resistant and its learned clusters (modeled by decision trees) are strongly interpretable and predictive. We systematically evaluate our algorithm using a common law dataset comprised of actual cases. Experimental results show that our algorithm significantly outperforms K-means and LDA on clustering dependency.