A continuous-time Markov process (CTMP) is a collection of variables indexed by a continuous quantity, time. It obeys the Markov property that the distribution over a future variable is independent of past variables given the state at the present time. We introduce continuous-time Markov process representations and algorithms for filtering, smoothing, expected sufficient statistics calculations, and model estimation, assuming no prior knowledge of continuous-time processes but some basic knowledge of probability and statistics. We begin by describing "flat" or unstructured Markov processes and then move to structured Markov processes (those arising from state spaces consisting of assignments to variables) including Kronecker, decision-diagram, and continuous-time Bayesian network representations. We provide the first connection between decision-diagrams and continuous-time Bayesian networks.

The recently proposed SPARse Factor Analysis (SPARFA) framework for personalized learning performs factor analysis on ordinal or binary-valued (e.g., correct/incorrect) graded learner responses to questions. The underlying factors are termed "concepts" (or knowledge components) and are used for learning analytics (LA), the estimation of learner concept-knowledge profiles, and for content analytics (CA), the estimation of question-concept associations and question difficulties. While SPARFA is a powerful tool for LA and CA, it requires a number of algorithm parameters (including the number of concepts), which are difficult to determine in practice. In this paper, we propose SPARFA-Lite, a convex optimization-based method for LA that builds on matrix completion, which only requires a single algorithm parameter and enables us to automatically identify the required number of concepts. Using a variety of educational datasets, we demonstrate that SPARFALite (i) achieves comparable performance in predicting unobserved learner responses to existing methods, including item response theory (IRT) and SPARFA, and (ii) is computationally more efficient.

Machine learning offers novel ways and means to design personalized learning systems wherein each student's educational experience is customized in real time depending on their background, learning goals, and performance to date. SPARse Factor Analysis (SPARFA) is a novel framework for machine learning-based learning analytics, which estimates a learner's knowledge of the concepts underlying a domain, and content analytics, which estimates the relationships among a collection of questions and those concepts. SPARFA jointly learns the associations among the questions and the concepts, learner concept knowledge profiles, and the underlying question difficulties, solely based on the correct/incorrect graded responses of a population of learners to a collection of questions. In this paper, we extend the SPARFA framework significantly to enable: (i) the analysis of graded responses on an ordinal scale (partial credit) rather than a binary scale (correct/incorrect); (ii) the exploitation of tags/labels for questions that partially describe the question-concept associations. The resulting Ordinal SPARFA-Tag framework greatly enhances the interpretability of the estimated concepts. We demonstrate using real educational data that Ordinal SPARFA-Tag outperforms both SPARFA and existing collaborative filtering techniques in predicting missing learner responses.

Robert Gagne (1968) defined a learning hierarchy as a set of specified intellectual capabilities or intellectual skills. The capabilities in the hierarchy have an ordered relationship to each other and the hierarchy, as a whole, bears some relation to a plan for effective instruction. The hierarchy is built in a manner to reflect that a lower level skill must be acquired or mastered before an upper-level one, that is, lower level capabilities are prerequisites for upper level ones. Intellectual capabilities or skills are the nodes of the hierarchy. Gagne (1968) defines them as cognitive strategies that denote capabilities for action. Additionally, they also depict a learning route, a path, from simple skills to a final complex capability. Learning hierarchies not only serve to represent effective instruction plans in terms of skills or capabilities, but also, they serve as diagnosis instruments for providing individual or personalized remediation to students. However, for classrooms with a large number of students, the application of learning hierarchies for individualized (remedial) instruction is a highly time consuming task. Learning hierarchies belong to the behaviorist view on cognition and www.ijera.com

This dissertation shows that careful injection of noise into sample data can substantially speed up Expectation-Maximization algorithms. Expectation-Maximization algorithms are a class of iterative algorithms for extracting maximum likelihood estimates from corrupted or incomplete data. The convergence speed-up is an example of a noise benefit or "stochastic resonance" in statistical signal processing. The dissertation presents derivations of sufficient conditions for such noise-benefits and demonstrates the speed-up in some ubiquitous signal-processing algorithms. These algorithms include parameter estimation for mixture models, the $k$-means clustering algorithm, the Baum-Welch algorithm for training hidden Markov models, and backpropagation for training feedforward artificial neural networks. This dissertation also analyses the effects of data and model corruption on the more general Bayesian inference estimation framework. The main finding is a theorem guaranteeing that uniform approximators for Bayesian model functions produce uniform approximators for the posterior pdf via Bayes theorem. This result also applies to hierarchical and multidimensional Bayesian models.

This paper establishes the consistency of a family of graph-cut-based algorithms for clustering of data clouds. We consider point clouds obtained as samples of a ground-truth measure. We investigate approaches to clustering based on minimizing objective functionals defined on proximity graphs of the given sample. Our focus is on functionals based on graph cuts like the Cheeger and ratio cuts. We show that minimizers of the these cuts converge as the sample size increases to a minimizer of a corresponding continuum cut (which partitions the ground truth measure). Moreover, we obtain sharp conditions on how the connectivity radius can be scaled with respect to the number of sample points for the consistency to hold. We provide results for two-way and for multiway cuts. Furthermore we provide numerical experiments that illustrate the results and explore the optimality of scaling in dimension two.

The agent is "always on," i.e. it is continuously available and aware (using a camera and infrared motion sensor) when the user is in its presence and can initiate interaction with the user, rather than requiring the user login to begin interaction. We expect that the agent will help reduce the user's isolation not just by always being around but also by specific activities that connect the user with friends, family and the local community. Our goal is for the agent to be a natural, humanlike presence that "resides" in the user's apartment. Beginning in the late summer of 2014, we will be placing our agents with users for a monthlong evaluation study. Figure 1: Virtual agent interface -- "Karen" Three issues of our project directly concern the topics of this workshop are: (1) the embodiment of the agent, (2) the engagement behaviors that are associated with being "always measures we will be using are questionnaires that assess the on," and (3) AI tools for support intelligent behavior.

We can also leverage medical ontologies/taxonomies to help Traditionally, artificial intelligence in medical applications abstract specific details to concepts that can be more easily has focused on improving the abilities of medical professionals introduced and then later refined when a patient is ready. Additionally, to perform tasks such as diagnosis (e.g., Shortliffe we can have annotations to provide information 1986; Wyatt and Spiegelhalter 1991; Garg et al. 2005; Vihinen about the authoritativeness of content. Furthermore, in many and Samarghitean 2008) or to aid in managing drug interactions cases information will need to travel beyond the patient to (e.g., Bindoff et al. 2007) or side effects (Edwards family or hired caregivers (Williams et al. 2002, p. 387), and Aronson 2000, p. 1258). These efforts target users who which means that multiple explanations will need to be generated have years of medical experience. In contrast, patients often based on the target individual's knowledge. Explanation have limited medical knowledge, and they may be coping generation also involves applications of user modeling with new life-threatening diagnoses that may require a number (e.g.

Computer Science researchers have long sought ways to apply the fruits of their labors to education. From the Logo turtles to the latest Cognitive Tutors, the allure of computers that will understand and help humans learn and grow has been a constant thread in Artificial Intelligence research. Now, advances in robotics and our understanding of Human-Robot Interaction make it feasible to develop physically-present robots that are capable of presenting educational material in an engaging manner, adapting online to sensory information from individual students, and building sophisticated, personalized models of a student’s mastery over complex educational domains. In this paper, we discuss how using physical robots as platforms for artificially intelligent tutors enables an expanded space of possible educational interactions. We also describe a work-in-progress to (1) extend previous work in personalized user models for robotic tutoring and (2) further explore the differences between interaction with physical robots and onscreen agents. Specifically, we are examining how embedding an tutoring interaction inside a story, game, or activity with an agent may differentially affect learning gains and engagement in interactions with physical robots and screen-based agents.

Engagement is a key factor in every social interaction, be it between humans or humans and robots. Many studies were aimed at designing robot behavior in order to sustain human engagement. Infants and children, however, learn how to engage their caregivers to receive more attention.We used a social robot platform, DragonBot, that learned which of its social behaviors retained human engagement. This was achieved by implementing a reinforcement learning algorithm, wherein the reward is the proximity and number of people near the robot. The experiment was run in the World Science Festival in New York, where hundreds of people interacted with the robot. After more than two continuous hours of interaction, the robot learned by itself that making a sad face was the most rewarding expression. Further analysis showed that after a sad face, people's engagement rose for thirty seconds. In other words, the robot learned by itself in two hours that almost no-one leaves a sad DragonBot.