Human-machine systems have expanded in terms of their sensing, communication, and computational capabilities. These capabilities have led to developments of a variety of sensor systems, like robotic platforms. There are benefits to these new sensor systems, however, these benefits have been offset by new difficulties; dynamic data overload, keeping pace with changing tempo, and managing data flows from multiple sensors feeds. One approach to manage data overload from multiple sensor feeds are computational models of attention. These models also address an important aspect of human-machine symbiosis, the need for machines agents to understand attention, manage interaction based on the flow of attention, and anticipate the flow of attention in the future. Unfortunately, existing computational models of attention use assumptions that limit their applicability to human-machine systems. The Artificial Attention Architecture is introduced and demonstrates how computational models of attention can be extended to handle multi-agent, multi-sensor systems. The Artificial Attention Architecture addresses important properties of human-machine systems like the need to build symbiosis between people searching for meaning in extensive data flows and the computational algorithms processing complex and dynamic data flows.

In this work, we introduce a technique that uses stories totrain virtual agents to exhibit believable behavior. This technique uses a compact representation of a story to define the space of acceptable behaviors and then uses this space to assign rewards to certain world states. We show the effectiveness of our technique with a case study in a modified gridworld environment called Pharmacy World. The results show that a reinforcement learning agent using Q-learning was able to learn a policy that results in believable behavior.

Scientific articles are not born equal. Some generate an entire discipline while others make relatively fewer contributions. When reviewing scientific literatures, it would be useful to identify those important articles and understand how they influence others. In this paper, we introduce J-Index, a quantitative metric modeling topic-level academic influence. J-Index is calculated based on the novelty of each article as well as its contributions to the articles where it is cited. We devise a generative model named Reference Topic Model (RefTM) which jointly utilizes the textual content and citation information in scientific literatures. We show how to learn RefTM to discover both the novelty of each paper and the strength of each citation. Experiments on a collection of more than 420,000 research papers demonstrate that RefTM outperforms the state-of-the-art approaches in terms of topic coherence as well as prediction performance, and validate J-Index's effectiveness of capturing topic-level academic influence in scientific literatures.

The problem of designing recommender systems for scholarly article citations has been actively researched with more than 200 publications appearing in the last two decades. In spite of this, no definitive results are available about what approaches work best. Arguably the most important reason for this lack of consensus is the dearth of standardised test collections and evaluation protocols, such as those provided by TREC-like forums. CiteSeerX, a "scholarly big dataset" has recently become available. However, this collection provides only the raw material that is yet to be moulded into Cranfield style test collections. In this paper, we discuss the limitations of test collections used in earlier work, and describe how we used CiteSeerX to design a test collection with a well-defined evaluation protocol. The collection consists of over 600,000 research papers and over 2,500 queries. We report some preliminary experimental results using this collection, which are indicative of the performance of elementary content-based techniques. These experiments also made us aware of some shortcomings of CiteSeerX itself.

In the past few years various government funding organizations such as the U.S. National Institutes of Health and the U.S.\ National Science Foundation have provided access to large publicly-available online databases documenting the grants that they have funded over the past few decades. These databases provide an excellent opportunity for the application of statistical text analysis techniques to infer useful quantitative information about how funding patterns have changed over time. In this paper we analyze data from the National Cancer Institute (part of National Institutes of Health) and show how text classification techniques provide a useful starting point for analyzing how funding for cancer research has evolved over the past 20 years in the United States.

The development of new scientific concepts today is an outcome of the accumulated knowledge built over time. Every scientific domain requires understanding of the trends of the dependencies between its subdomains. Analyses of trends to capture such dependencies using conventional document modeling techniques is a challenging task due to two reasons: (1) conventional vector-space modeling based representation of documents does not realize the history of the content, and (2) neither feature-level nor document-level causality is provided with any digital library metadata or citation network. In this paper, we propose an intuitive temporal representation of a scientific article that encodes inherent historic characteristics of the content. This intuitive representation of each document is then leveraged to discover causal relationships between scientific articles. In addition, we provide a mechanism to explore the lineage of each document in terms of other previously published documents, which illustrates how the theme of the document under analysis evolved over time. Empirical studies reported in the paper show that the proposed technique identifies meaningful causal relationships and discovers meaningful lineage in the scientific literature that could not be discovered through the citation network of the articles.

Retrieval systems for scholarly literature offer the ability for the scientific community to search, explore and download scholarly articles across various scientific disciplines. Mostly used by the experts in the particular field, these systems contain user community logs including information on user specific downloaded articles. In this paper we present a novel approach for automatically evaluating document similarity models in large collections of scholarly publications. Unlike typical evaluation settings that use test collections consisting of query documents and human annotated relevance judgments, we use download logs to automatically generate pseudo-relevant set of similar document pairs. More specifically we show that consecutively downloaded document pairs, extracted from a scholarly information retrieval (IR) system, could be utilized as a test collection for evaluating document similarity models. Another novel aspect of our approach lies in the method that we employ for evaluating the performance of the model by comparing the distribution of consecutively downloaded document pairs and random document pairs in log space. Across two families of similarity models, that represent documents in the term vector and topic spaces, we show that our evaluation approach achieves very high correlation with traditional performance metrics such as Mean Average Precision (MAP), while being more efficient to compute.

An effective approach to solving problems involving mixed (continuous and discrete) variables and constraints, such as hybrid systems, is to decompose them into subproblems and integrate dedicated solvers geared toward those subproblems. Here, we introduce a new framework based on a tree search algorithm to solve hybrid discrete-continuous problems that incorporates: (1) a quantum annealer that samples from the configuration space for the discrete portion and provides information about the quality of the samples, and (2) a classical computer that makes use of information from the quantum annealer to prune and focus the search as well as check a continuous constraint. We consider four variants of our algorithm, each with progressively more guidance from the results provided by the quantum annealer. We empirically test our algorithm and compare the variants on a simplified Mars Lander task scheduling problem. Variants with more guidance from the quantum annealer have better performance.

Planning in hybrid systems is important for dealing with real world applications. PDDL+ supports this representation of domains with mixed discrete and continuous dynamics, and supports events and processes modeling exogenous change. Motivated by numerous SAT-based planning approaches, we propose an approach to PDDL+ planning through SMT, describing an SMT encoding that captures all the features of the PDDL+ problem as published by Fox and Long (2006). The encoding can be applied on domains with nonlinear continuous change. We apply this encoding in a simple planning algorithm, demonstrating excellent results on a set of benchmark problems.

We present a new game-theoretic framework where Bayesian players engage in a Markov game and each has private but imperfect information regarding other players' types. Instead of utilizing Harsanyi's abstract types and a common prior distribution, we construct player types whose structure is explicit and induces a finite level belief hierarchy. We characterize equilibria in this game and formalize the computation of finding such equilibria as a constraint satisfaction problem. The effectiveness of the new framework is demonstrated on two ad hoc team work domains.