Attackers use various domain squatting techniques to convince users that their services are legitimate. Previous work has shown that methods liketyposquatting, where single characters are removed or duplicated, can successfully deceive users.

Individual and organizational computer security rests on how people interpret and use the security information they are presented. One challenge is determining whether a given URL is safe or not. This paper explores the visual behaviors that users employ to gauge URL safety. We conducted a user study on 20 participants wherein participants classified URLs as safe or unsafe while wearing an eye tracker that recorded eye gaze (where they look) and pupil dilation (a proxy for cognitive effort). Among other things, our findings suggest that: users have a cap on the amount of cognitive resources they are willing to expend on vetting a URL; they tend to believe that the presence of www in the domain name indicates that the URL is safe; and they do not carefully parse the URL beyond what they perceive as the domain name.

Human error is one of the most common causes of vul- nerability in a secure system. However it is often overlooked when these systems are tested, partly because human tests are costly and very hard to repeat. We have developed a community of agents that test secure systems by running standard windows software while performing collaborative group tasks, mimicking more realistic patterns of communication and traffic, as well as human fatigue and errors. This system is being deployed on a large cyber testing range. One key attribute of humans is flexibility of response in order to achieve their goals when unexpected events occur. Our agents use reactive planning within a BDI architecture to flexibly replan if needed. Since the agents are goal-oriented, we are able to measure the impact of cyber attacks on mission accomplishment, a more salient measure of protection than raw penetration. We show experimentally how the agent teams can be resilient under attacks that are partly successful, and also how an organizational structure can lead to emergent properties of the traffic in the network.

Attacks on computer systems are rapidly becoming more numerous and more sophisticated, and current preventive techniques do not seem able to keep pace. Many successful attacks can be attributed to user errors: for example, while focused on other tasks, users may succumb to 'social engineering' attacks such as phishing or trojan horses. Warnings about the danger of these attacks are often vaguely worded and given long before the dangers are realized, and are therefore too easy to ignore. However, we hypothesize that users are more likely to be persuaded by messages that (1) leverage mental models to describe the dangers, (2) describe particular vulnerabilities that the user may be exposed to and (3) are delivered close in time before the danger may actually be realized. We discuss the design and initial implementation of a system to achieve this. It first shows a video about a potential danger, then creates warnings tailored to the user's environment and given at the time they may be most useful, displaying a still frame or snippet from the video to remind the user of the potential danger. The system uses templates of user activities as input to a markov logic network to recognize potentially risky behaviors. This approach can identify likely next steps that can be used to predict immediate danger and customize warnings.

Numerous techniques exist to help users automate repetitive tasks; however, none of these methods fully support end-user creation, use, and modification of the learned tasks. We present an integrated task learning system (ITL) that learns executable procedures based on user demonstration and instruction, constituting a first step toward a broader solution for procedure management. We discuss our deployment of ITL into a collaborative command-and-control system. In this complex domain, ITL's performance with end users doing real tasks indicates that providing multiple, integrated learning techniques both extends functionality and improves user experience. Our experience in integrat-ing this system also provides key insights for future designs of domain-independent task learning systems, specifically in supporting users' ability to understand and edit lengthy procedures.

To control intelligent tools that perform a variety of complex procedures, users need to be able to both modify existing procedure descriptions and communicate new procedures. In one approach, the user describes fragments of a procedure with text, and the tool searches the space of potential procedures for a match. This approach sometimes provides too little guidance for users, yet providing templates for guidance can require an expensive knowledge engineering effort in each new domain. We investigate the use of case-based reasoning to help guide the user, treating previously-defined procedures in the domain as cases. We describe domainindependent methods to find similar procedures while the user creates or modifies a procedure, to suggest potential steps to copy and to manage mapping the variables from the existing procedure into the procedure being edited. In some cases, the mapping tool suggests auxiliary steps to copy along with the desired steps, following an approach similar to derivational analogy. We evaluate the potential of this approach with an implemented tool, CB-Tailor, in a travel domain containing a number of procedures that may be added by the user. Our experiences suggest that the tool can provide useful guidance in a realistic set of situations.

Many real-world domains resist analysis because they are best characterized by a variety of data types, including relational, spatial, and temporal components. Examples of such domains include disease outbreaks, criminal networks, and the World-Wide Web. We present two types of visualizations based on physical metaphors that facilitate fusion, analysis, and deep understanding of relational, spatio-temporal data. The first visualization is based on the metaphor of fluid flow through elastic pipes, and the second on wave propagation. We discuss both types of visualizations in the context of fusing information about the activities of scientists over time with the goal of constructing career histories

Grid applications require allocating a large number of heterogeneous tasks to distributed resources. A good allocation is critical for efficient execution. However, many existing grid toolkits use matchmaking strategies that do not consider overall efficiency for the set of tasks to be run. We identify two families of resource allocation algorithms: task-based algorithms, that greedily allocate tasks to resources, and workflow-based algorithms, that search for an efficient allocation for the entire workflow. We compare the behavior of workflow-based algorithms and task-based algorithms, using simulations of workflows drawn from a real application and with varying ratios of computation cost to data transfer cost. We observe that workflow-based approaches have a potential to work better for data-intensive applications even when estimates about future tasks are inaccurate.

In order for intelligent systems to be applicable in a wide range of situations, end users must be able to modify their task descriptions. We introduce Tailor, a system that allows users to modify task information through instruction. In this approach, the user enters a short sentence to describe the desired change. The system maps the sentence into valid, plausible modifications and checks for unexpected side-effects they may have, working interactively with the user throughout the process. We conducted preliminary tests in which subjects used Tailor to make modifications to domains drawn from the eHow website, applying modifications posted by readers as 'tips'. In this way the subjects acted as interpreters between Tailor and the human-generated descriptions of modifications. Almost all the subjects were able to make all modifications to the process descriptions with Tailor, indicating that the interpreter role is quite natural for users.

The PRODIGY system is based on bidirectional planning, which is a combination of goal-directed reasoning with simulated execution. Researchers have implemented a series of planners that utilize this search strategy, and demonstrated that it is an efficient technique, a fair match to other successful planners; however, they have provided few formal results on the common principles underlying the developed algorithms. We formalize bidirectional planning, elucidate some techniques for improving its efficiency and show how different strategies for controlling search complexity give rise to different versions of PRODIGY. In particular, we demonstrate that PRODIGY is not complete and discuss advantages and drawbacks of its incompleteness. We then develop a complete bidirectional planner and compare it experimentally with PRODIGY. We show that the complete planner is almost as fast as PRODIGY and solves a wider range of problems.

For the manual semantic markup of documents to become wide-spread, usersmust be able to express annotations that conform to ontologies (orschemas) that have shared meaning. However, a typical user is unlikelyto be familiar with the details of the terms as defined by the ontology authors. In addition, the idea to be expressed may not fit perfectly within a pre-defined ontology. The ideal tool should help users find apartial formalization that closely follows the ontology where possiblebut deviates from the formal representation where needed. We describe animplemented approach to help users create semi-structured semantic annotations for a document according to an extensible OWL ontology. In our approach, users enter a short sentence in free text to describe allor part of a document, and the system presents a set of potential paraphrases of the sentence that are generated from valid expressions inthe ontology, from which the user chooses the closest match. We use a combination of off-the-shelf parsing tools and breadth-first search of expressions in the ontology to help users create valid annotations starting from free text. The user can also define new terms to augmentthe ontology, so the potential matches can improve over time.

Interactive tools to help users author plans or processes are essential in a variety of domains. KANAL helps users author sound plans by simulating them, checking for a variety of errors and presenting the results in an accessible format that allows the user to see an overview of the plan steps or timelines of objects in the plan. From our experience in two domains, users tend to interleave plan authoring and plan checking while extending background knowledge of actions. This has led us to refine KANAL to provide a high-level overview of plans and integrate a tool for refining the background knowledge about actions used to check plans. We report on these lessons learned and new directions in KANAL.

Eliciting complex logical rules directly from logic-naive subject matter experts (SMEs) is a challenging knowledge capture task. We describe a large-scale experiment to evaluate tools designed to produce SME-authored rule bases. We assess the quality of the rule bases with respect to the: 1) performance on the addressed functional task (military course of action (COA) critiquing); and 2) intrinsic knowledge representation quality. In the course of this assessment, we note both strengths and weaknesses in the state of the art, and accordingly suggest some foci for future development in this important technology area.

This paper describes an integrated acquisition interface that includes several techniques previously developed to support users in various ways as they add new knowledge to an intelligent system. As a result of this integration, the individual techniques can take better advantage of the context in which they are invoked and provide stronger guidance to users. We describe the current implementation using examples from a travel planning domain, and demonstrate how users can add complex knowledge to the system.

Planning is a complex reasoning task that is well suited for the study of improving performance and knowledge by learning, i.e. by accumulation and interpretation of planning experience. PRODIGY is an architecture that integrates planning with multiple learning mechanisms. Learning occurs at the planner's decision points and integration in PRODIGY is achieved via mutually interpretable knowledge structures. This article describes the PRODIGY planner, briefly reports on several learning modules developed earlier along the project, and presents in more detail two recently explored methods to learn to generate plans of better quality. We introduce the techniques, illustrate them with comprehensive examples, and show preliminary empirical results. The article also includes a retrospective discussion of the characteristics of the overall PRODIGY architecture and discusses their evolution within the goal of the project of building a large and robust integrated planning and learning system.

We describe our experience with the object-oriented design methodology OMT [19] applied to a project in an undergraduate software engineering course at Carnegie Mellon University.The project involved 30 students previously unfamiliar with objectoriented modeling.They designed and implemented a system of 125 classes (27,000 lines of C++ and C code) in the relatively short time of 15 weeks.We describe the overall structure of the project and the system model, the usefulness of OMT and its impact on communication, and discuss some of the problems encountered during the development of the system.