Many intelligent tutoring system (ITS) researchers are looking at ways to detect and to respond to student emotional states (for instance animated pedagogical agents that mirror student emotion). Such interventions are complicated to build, and do not take advantage of the potential for teachers to be part of the process. We present two studies that intervene when a student is having trouble by presenting the student with a YouTube video that is recorded by their own teacher and that delivers a motivational message to help them to persist with the learning session. We experimentally compared two different motivational interventions, which are both grounded in the literature on student affect and motivation. We also had a control condition that had no video. We found that when looking at students’ self-reports on the value of mathematics, we found a main effect of condition for the value-video. In Study 2 we examined whether these 60-second videos could impact homework completion rates and found that in fact homework completion rates were higher for students in the value-video condition. The present research is suggestive of a somewhat novel use of teacher-generated content that could easily be incorporated into other ITSs.

Classified ad sites routinely process hundreds of thousands to millions of posted ads, and only a small percentage of those may be fraudulent. Online scammers often go through a great amount of effort to make their listings look legitimate. Examples include copying existing advertisements from other services, tunneling through local proxies, and even paying for extra services using stolen account information. This paper focuses on applying knowledge discovery concepts towards the detection of online, classified fraud. Traditional data mining is used to extract relevant attributes from an online classified advertisements database and machine learning algorithms are applied to discover patterns and relationships of fraudulent activity. With our proposed approach, we will demonstrate the effectiveness of applying data mining techniques towards the detection of fraud in online classified advertisements.

Research on individual differences indicates that students vary in how they interact with and perform while using intelligent tutoring systems (ITSs). However, less research has investigated how individual differences affect students’ interactions with game-based features. This study examines how learning outcomes and interactions with specific game-based features (off-task personalization vs. on-task mini games) within a game-based ITS, iSTART-ME, vary as a function of students’ performance orientation. The current study (n=40) is part of a larger study (n=126) conducted with high school students. The analyses in this study focus on those students assigned to iSTART-ME. Results indicate that students with higher levels of performance orientation perform better during training, progress further within the system, and interact less frequently with off-task game-based features. These results provide further evidence that individual differences play an important role in influencing students’ interactions and achievement within learning environments.

In student modeling, the concept of "mastery learning" i.e. that a student continues to learn a skill till mastery is attained is important. Usually, mastery is defined in terms of most recent student performance. This is also the case with models such as Knowledge Tracing which estimate knowledge solely based on patterns of questions a student gets correct and the task usually is to predict immediate next action of the student. In retrospect however, it is not clear if this is a good definition of mastery since it is perhaps more useful to focus more on student retention over a longer period of time. This paper improves a recently introduced model by Wang and Beck that predicts long term student performance by clustering the students and generating multiple predictions by using a recently developed ensemble technique. Another contribution is that we introduce a novel clustering algorithm we call "Regularity Clustering" and show that it is superior in the task of predicting student retention over more popular techniques such as k-means and Spectral Clustering.

Introduction to the papers presented at the 2013 FLAIRS Conference Special Track on Intelligent Tutoring Systems

Educators who wish to engage their students in multi-agent experimentation and learning need an inexpensive multi-robot system that leverages existing equipment and open-source software. This paper proposes Overwatch as an inexpensive educational tool for teaching and experimenting in multi-robot systems. The interaction of multiple agents within a single environment is an important area of study. It is vital that agents within the environment perceive other agents as intelligent, acting within the environment as cooperative teammates or as competitive members of another team. To do so, the system must meet three goals: first, to allow multiple robots to communicate and coordinate; second, to localize within a shared global coordinate system; third, to recognize their teammates and other teams. The cost and scale of such experimental platforms places them outside the reach of many educational institutions or limits the number of agents that are interacting within the system \cite{Liu201160}. The goal of Overwatch is to create an experimental platform for multi-agent systems that is comprised of much smaller, albeit less capable, robots, many of which are prevalent in academic institutions already. Making use of available open-source libraries and utilizing lower cost robots, such as Scribblers, allows for experiments with many agents. This enables Overwatch to fit into the budget limitations of an academic setting. The Overwatch platform provides the Scribblers with global localization capabilities. This paper presents the system in detail and includes experiments to show its ability to localize, interact with other agents, and coordinate behaviors with these other agents. Additionally, the details to setup this system are also included.

This study expands upon an existing model of students’ reading comprehension ability within an intelligent tutoring system. The current system evaluates students’ natural language input using a local student model. We examine the potential to expand this model by assessing the linguistic features of self-explanations aggregated across entire passages. We assessed the relationship between 126 students’ reading comprehension ability and the cohesion of their aggregated self-explanations with three linguistic features. Results indicated that the three cohesion indices accounted for variance in reading ability over and above the features used in the current algorithm. These results demonstrate that broadening the window of NLP analyses can strengthen student models within ITSs.

This study compares automated scoring increases and linguistic changes for student writers in two groups: a group that used an intelligent tutoring system embedded with an automated writing evaluation component (Writing Pal) and a group that used only the automated writing evaluation component. The primary goal is to examine automated scoring differences in both groups from pretest to posttest essays to investigate score gains and linguistic development. The study finds that both groups show significant increases in automated writing scores and significant development in lexical, syntactic, cohesion, and rhetorical features. However, the Writing-Pal group shows greater raw frequency gains (i.e., negative v. positive gains).

In this paper, we study optimization methods consisting of iteratively minimizing surrogates of an objective function. By proposing several algorithmic variants and simple convergence analyses, we make two main contributions. First, we provide a unified viewpoint for several first-order optimization techniques such as accelerated proximal gradient, block coordinate descent, or Frank-Wolfe algorithms. Second, we introduce a new incremental scheme that experimentally matches or outperforms state-of-the-art solvers for large-scale optimization problems typically arising in machine learning.

In an online contract selection problem there is a seller which offers a set of contracts to sequentially arriving buyers whose types are drawn from an unknown distribution. If there exists a profitable contract for the buyer in the offered set, i.e., a contract with payoff higher than the payoff of not accepting any contracts, the buyer chooses the contract that maximizes its payoff. In this paper we consider the online contract selection problem to maximize the sellers profit. Assuming that a structural property called ordered preferences holds for the buyer's payoff function, we propose online learning algorithms that have sub-linear regret with respect to the best set of contracts given the distribution over the buyer's type. This problem has many applications including spectrum contracts, wireless service provider data plans and recommendation systems.