Introductory engineering courses within large universities often have annual enrollments which can reach up to a thousand students. It is very challenging to achieve differentiated instruction in classrooms with class sizes and student diversity of such great magnitude. Professors can only assess whether students have mastered a concept by using multiple choice questions, while detailed homework assignments, such as planar truss diagrams, are rarely assigned because professors and teaching assistants would be too overburdened with grading to return assignments with valuable feedback in a timely manner. In this paper, we introduce Mechanix, a sketch-based deployed tutoring system for engineering students enrolled in statics courses. Our system not only allows students to enter planar truss and free body diagrams into the system just as they would with pencil and paper, but our system checks the student's work against a hand-drawn answer entered by the instructor, and then returns immediate and detailed feedback to the student. Students are allowed to correct any errors in their work and resubmit until the entire content is correct and thus all of the objectives are learned. Since Mechanix facilitates the grading and feedback processes, instructors are now able to assign free response questions, increasing teacher's knowledge of student comprehension. Furthermore, the iterative correction process allows students to learn during a test, rather than simply displaying memorized information.

This case study concerns the use of software agent advisors to improve efficiency and quality in issue tracking activities of development teams at the world's largest medical device manufacturer. Each software agent monitors, interacts with, and learns from its environment and user, recognizing when and how to provide different kinds of advice and support to facilitate issue tracking activities without directly modifying anything or otherwise violating domain constraints. The deployed software agent has not only enjoyed regular and growing use, but contributed to significant improvements. Issue rejection was significantly reduced and more focused, yielding significant quality and efficiency gains such as fewer reviews by quality assurance. This success reflects the benefits of the underlying AI technology.

With rapid profusion of video data, automated surveillanceand intrusion detection is becoming closer to reality. In orderto provide timely responses while limiting false alarms, an intrusiondetection system must balance resources (e.g., time)and accuracy. In this paper, we show how such a system canbe modeled with a partially observable Markov decision process(POMDP), representing possible computer vision filtersand their costs in a way that is similar to human vision systems.The POMDP representation can be optimized to producea dynamic sequence of operations and achieve a tradeoffbetween time and detection quality, taking into accountuncertainty in the filter predictions. In a set of experiments onactual video data, we show that our method can both outperformstatic “expert” models and scale to large dynamic domains.These results suggest that our method could be usedin real-world intrusion detection systems.

Spoken dialogue is increasingly central to systems that assist people. As the tasks that people and machines speak about together become more complex, however, users’ dissatisfaction with those systems is an important concern. This paper presents a novel approach to learning for spoken dialogue systems. It describes embedded wizardry, a methodology for learning from skilled people, and applies it to a library whose patrons order books by telephone. To address the challenges inherent in this application, we introduce RFW+, a domain-independent, feature-selection method that considers feature categories. Models learned with RFW+ on embedded-wizard data improve the performance of a traditional spoken dialogue system.

This paper shows how to improve the recommendations of an interaction-based collaborative filtering (IBCF) recommender used in online dating. Previous work has shown that IBCF works well in this domain, although it tends to rank popular candidates highly, which leads to these users receiving a large number of contacts. We address this problem by using a Decision Tree model as a "critic" to re-rank the candidates generated by IBCF, effectively promoting less popular candidates. This method was first evaluated on historical data from a large online dating site and then trialled live on the same site by providing recommendations to a large number of users throughout a 9 week period. The live trial confirmed the consistency of the analysis on historical data and the ability of the method to generate suitable candidates over an extended period. Our recommendations gave higher success rates than those for a control group made with a baseline recommender.

This paper introduces a new technique for intelligently exploring the reachability set of a spacecraft: the set of trajectories from a given initial condition that are possible under a specified range of control actions. The high dimension of this problem and the nonlinear nature of gravitational interactions make the geometry of these sets complicated, hard to compute, and all but impossible to visualize. Currently, exploration of a problem’s state space is done heuristically, based on previously identified solutions. This potentially misses out on improved mission design solutions that are not close to previous approaches. The goal of the work described here is to map out reachability sets automatically. This would not only aid human mission planners, but also allow a spacecraft to determine its own course without input from Earth-based controllers. Brute-force approaches to this are computationally prohibitive, so one must focus the effort on regions that are of interest: where neighboring trajectories diverge quickly, for instance, or come close to a body that the spacecraft is orbiting. In this paper, we focus on the first of those two criteria; the goal is to identify regions in the system’s state space where small changes have large effects— or vice versa—and concentrate the computational mesh accordingly.

We have developed a method for statistical anomaly detection which has been deployed in a tool for condition monitoring of train fleets. The tool is currently used by several railway operators over the world to inspect and visualize the occurrence of event messages generated on the trains. The anomaly detection component helps the operators to quickly find significant deviations from normal behavior and to detect early indications for possible problems. The savings in maintenance costs comes mainly from avoiding costly breakdowns, and have been estimated to several million Euros per year for the tool. In the long run, it is expected that maintenance costs can be reduced with between 5 and 10 % by using the tool.

We investigate machine learning methods for diagnostic screening of heart disease. Coronary heart disease is the leading cause of death in the US, causing more deaths than all types of cancers combined. Early diagnosis of heart disease in women is harder than it is in men and typically requires the administration of several clinical tests on the patient. Most risk stratification methods aggregate the results of such tests, including the risky, invasive procedures that cannot be administered on all patients. In this paper, our goal is to identify patients who are under high-risk of having heart disease and related adverse events, using a minimal number of diagnostic tests, especially less invasive ones. The low frequency of patients with severe heart disease in the dataset is challenging for most conventional machine learning methods. To overcome this problem, we develop and apply a cost-sensitive k nearest neighbor algorithm. Our contributions are two fold: First, we compare the predictive value of several diagnostic procedures for heart disease, including electrocardiography, angiography, radionuclide perfusion and conclude that in womens heart disease, certain combinations of non-invasive techniques are more predictive than some of the widely used invasive procedures. Then, we evaluate held out data and achieve an AUROC over 0.70, signifying valuable clinical utility, using only the least costly and least invasive tests.

Such systems humans are much more efficient at abstracting and matching take advantage of human abilities--particularly in vision, visual cues across piece borders based on their content. For natural language, and pattern recognition--to handle example, a person looking at a piece of a shredded document instances and aspects of problems that are difficult for can recognize a letter that is only partially present, and an computers. The ESP game (von Ahn and Dabbish 2008), experienced archaeologist looking at a particular piece of FoldIt (Cooper et al. 2010), and reCAPTCHA (von Ahn et a broken artifact can recognize unique patterns that extend al. 2008) are a few examples of successful systems that draw beyond the fragment. Unfortunately, for a human to find a on human contributors and machine computations to tackle matching piece still requires scanning through the pieces, problems in image labeling, protein folding, and text digitization.

The average home computer user needs help in reducing the security risk of their home computer. We are working on an alternative approach from current home security software in which a software agent helps a user manage his/her security risk. Planning is integral to the design of this agent in several ways. First, planning can be used to make the underlying security model manageable by generating attack paths to identify vulnerabilities that are not a problem for a particular user/home computer. Second, planning can be used to identify interventions that can either avoid the vulnerability or mitigate the damage should it occur. In both cases, a central capability is that of generating alternative plans so as to find as many possible ways to trigger the vulnerability and to provide the user with options should the obvious not be acceptable. We describe our security model and our state-based approach to generating alternative plans. We show that the state-based approach can generate more diverse plans than a heuristic-based approach. However, the state-based approach sometimes generates this diversity with better quality at higher search cost.