Smartphones incorporate many diverse and powerful sensors, which creates exciting new opportunities for data mining and human-computer interaction. In this paper we show how standard classification algorithms can use labeled smartphone-based accelerometer data to identify the physical activity a user is performing. Our main focus is on evaluating the relative performance of impersonal and personal activity recognition models. Our impersonal (i.e., universal) models are built using training data from a panel of users and are then applied to new users, while our personal models are built with data from each user and then applied only to new data from that user. Our results indicate that the personal models perform dramatically better than the impersonal models—even when trained from only a few minutes worth of data. These personal models typically even outperform hybrid models that utilize both personal and impersonal data. These results strongly argue for the construction of personal models whenever possible. Our research means that we can unobtrusively gain useful knowledge about the habits of potentially millions of users. It also means that we can facilitate human computer interaction by enabling the smartphone to consider context and this can lead to new and more effective applications.

Knowledge workers work on many different tasks and must often switch between those tasks. In earlier work, we have shown the benefits of automatically capturing contexts for tasks for a specific category of knowledge worker, software programmers. Captured contexts facilitate task switches and reduce information overload by enabling the display of only the information relevant to the task-at-hand. In this paper, we describe the results of two studies of the use of captured contexts for a broad range of knowledge workers. The first study we describe is a field study of eight knowledge workers who used the model in their daily work for up to 25 days on tasks involving both file and web documents. We found that these knowledge workers need information to decay from their context and that our model is adequate at automatically trimming contexts. The second study is a case study of the use of contexts to support the operations of a software development company. We analyzed task contexts from hundreds of days of work from three users and found similar trends of information decaying from contexts. Results from each study also shed more light on the nature of mixed artifact task contexts.

We introduce a laboratory text on environmental and societal sustainability applications that can be a supplemental resource for any undergraduate AI course. The lab text, entitled Artificial Intelligence for Computational Sustainability: A Lab Companion, is brand new and incomplete; freely available through Wikibooks; and open to community additions of projects, assignments, and explanatory material on AI for sustainability. The project adds to existing educational efforts of the computational sustainability community, encouraging the flow of knowledge from research to education and public outreach. Besides summarizing the laboratory book, this paper touches on its implications for integration of research and education, for communicating science to the public, and other broader impacts.

In recent years, major research advances have taken place in the intersection of computer science and economics, but this material has so far been taught primarily at the graduate level. This paper describes a novel semester-long undergraduate-level course in the intersection of computer science and economics at Duke University, titled “CPS 173: Computational Microeconomics.”

As growing numbers of real world applications involve imbalanced class distribution or unequal costs for mis- classification errors in different classes, learning from imbalanced class distribution is considered to be one of the most challenging issues in data mining research. This study empirically investigates the sensitivity of bagging predictors with respect to 12 algorithms and 9 levels of class distribution on 14 imbalanced data-sets by using statistical and graphical methods to address the important issue of understanding the effect of vary- ing levels of class distribution on bagging predictors. The experimental results demonstrate that bagging NB and MLP are insensitive to various levels of imbalanced class distribution.

In case of a plan failure, plan-repair is a more promising solution than replanning from scratch. The effectiveness of plan-repair depends on knowledge of which plan action failed and why. Therefore, in this paper, we propose an Extended Spectrum Based Diagnosis approach that efficiently pinpoints failed actions. Unlike Model Based Diagnosis (MBD), it does not require the fault models and behavioral descriptions of actions. Our approach first computes the likelihood of an action being faulty and subsequently proposes optimal probe locations to refine the diagnosis. We also exploit knowledge of plan steps that are instances of the same plan operator to optimize the selection of the most informative diagnostic probes. In this paper, we only focus on diagnostic aspect of plan-repair process.

Computing similarity between short microblogs is an important step in microblog recommendation. In this paper, we investigate a topic based approach and a WordNet based approach to estimate similarity scores between microblogs and recommend top related ones to users. Empirical study is conducted to compare their recommendation effectiveness using two evaluation measures. The results show that the WordNet based approach has relatively higher precision than that of the topic based approach using 548 tweets as dataset. In addition, the Kendall tau distance between two lists recommended by WordNet and topic approaches is calculated. Its average of all the 548 pair lists tells us the two approaches have the relative high disaccord in the ranking of related tweets.

This paper addresses the problem of automated advice provision in settings that involve repeated interactions between people and computer agents. This problem arises in many real world applications such as route selection systems and office assistants. To succeed in such settings agents must reason about how their actions in the present influence people's future actions. The paper describes several possible models of human behavior that were inspired by behavioral economic theories of people's play in repeated interactions. These models were incorporated into several agent designs to repeatedly generate offers to people playing the game. These agents were evaluated in extensive empirical investigations including hundreds of subjects that interacted with computers in different choice selections processes. The results revealed that an agent that combined a hyperbolic discounting model of human behavior with a social utility function was able to outperform alternative agent designs. We show that this approach was able to generalize to new people as well as choice selection processes that were not used for training. Our results demonstrate that combining computational approaches with behavioral economics models of people in repeated interactions facilitates the design of advice provision strategies for a large class of real-world settings.

Building upon previous security applications of computational game theory, this paper presents PROTECT, a game-theoretic system deployed by the United States Coast Guard (USCG) in the port of Boston for scheduling their patrols. USCG has termed the deployment of PROTECT in Boston a success, and efforts are underway to test it in the port of New York, with the potential for nationwide deployment. PROTECT is premised on an attacker-defender Stackelberg game model and offers five key innovations. First, this system is a departure from the assumption of perfect adversary rationality noted in previous work, relying instead on a quantal response (QR) model of the adversary's behavior - to the best of our knowledge, this is the first real-world deployment of the QR model. Second, to improve PROTECT's efficiency, we generate a compact representation of the defender's strategy space, exploiting equivalence and dominance. Third, we show how to practically model a real maritime patrolling problem as a Stackelberg game. Fourth, our experimental results illustrate that PROTECT's QR model more robustly handles real-world uncertainties than a perfect rationality model. Finally, in evaluating PROTECT, this paper provides real-world data: (i) comparison of human-generated vs PROTECT security schedules, and (ii) results from an Adversarial Perspective Team's (human mock attackers) analysis.

We propose to model relative attributes that capture the relationships between images and objects in terms of human-nameable visual properties. For example, the models can capture that animal A is 'furrier' than animal B, or image X is 'brighter' than image B. Given training data stating how object/scene categories relate according to different attributes, we learn a ranking function per attribute. The learned ranking functions predict the relative strength of each property in novel images. We show how these relative attribute predictions enable a variety of novel applications, including zero-shot learning from relative comparisons, automatic image description, image search with interactive feedback, and active learning of discriminative classifiers. We overview results demonstrating these applications with images of faces and natural scenes. Overall, we find that relative attributes enhance the precision of communication between humans and computer vision algorithms, providing the richer language needed to fluidly "teach" a system about visual concepts.