A person seeking someone else's attention is normally able to quickly assess how interruptible they are. This assessment allows for behavior we perceive as natural, socially appropriate, or simply polite. On the other hand, today's computer systems are almost entirely oblivious to the human world they operate in, and typically have no way to take into account the interruptibility of the user. This paper presents a Wizard of Oz study exploring whether, and how, robust sensor-based predictions of interruptibility might be constructed, which sensors might be most useful to such predictions, and how simple such sensors might be. The study simulates a range of possible sensors through human coding of audio and video recordings. Experience sampling is used to simultaneously collect randomly distributed self-reports of interruptibility. Based on these simulated sensors, we construct statistical models predicting human interruptibility and compare their predictions with the collected self-report data. The results of these models, although covering a demographically limited sample, are very promising, with the overall accuracy of several models reaching about 78%. Additionally, a model tuned to avoiding unwanted interruptions does so for 90% of its predictions, while retaining 75% overall accuracy.

Interest has been growing in opportunities to build and deploy statistical models that can infer a computer user’s current interruptability from computer activity and relevant contextual information. We describe a system that intermittently asks users to assess their perceived interruptability during a training phase and that builds decision-theoretic models with the ability to predict the cost of interrupting the user. The models are used at run-time to compute the expected cost of interruptions, providing a mediator for incoming notifications, based on a consideration of a user’s current and recent history of computer activity, meeting status, location, time of day, and whether a conversation is detected.

Although electronic communication plays an important role in the modern workplace, the interruptions created by poorly-timed attempts to communicate are disruptive. Prior work suggests that sharing an indication that a person is currently busy might help to prevent such interruptions, because people could wait for a person to become available before attempting to initiate communication. We present a context-aware communication client that uses the built-in microphones of laptop computers to sense nearby speech. Combining this speech detection sensor data with location, computer, and calendar information, our system models availability for communication, a concept that is distinct from the notion of presence found in widely-used systems. In a four week study of the system with 26 people, we examined the use of this additional context. To our knowledge, this is the first field study to quantitatively examine how people use automatically sensed context and availability information to make decisions about when and how to communicate with colleagues. Participants appear to have used the provided context to as an indication of presence, rather than considering availability. Our results raise the interesting question of whether sharing an indication that a person is currently unavailable will actually reduce inappropriate interruptions.

The computer and communication systems that office workers currently use tend to interrupt at inappropriate times or unduly demand attention because they have no way to determine when an interruption is appropriate. Sensor-based statistical models of human interruptibility offer a potential solution to this problem. Prior work to examine such models has primarily reported results related to social engagement, but it seems that task engagement is also important. Using an approach developed in our prior work on sensor-based statistical models of human interruptibility, we examine task engagement by studying programmers working on a realistic programming task. After examining many potential sensors, we implement a system to log low-level input events in a development environment. We then automatically extract features from these low-level event logs and build a statistical model of interruptibility. By correctly identifying situations in which programmers are non-interruptible and minimizing cases where the model incorrectly estimates that a programmer is non-interruptible, we can support a reduction in costly interruptions while still allowing systems to convey notifications in a timely manner. Author Keywords Situationally appropriate interaction, managing human attention, sensor-based interfaces, context-aware computing, machine learning, interruptibility.

For the estimation of user interruptability in wearable and mobile settings, we propose in [8] to distinguish between the users ’ personal and social interruptability. In this paper, we verify this thesis with a user study on 24 subjects. Results show that there is a significant difference between social and personal interruptability. Further, we present a novel approach to estimate the social and personal interruptability of a user from wearable sensors. It is scalable for a large number of sensors, contexts, and situations and allows for online adaptation during run-time. We have developed a wearable platform, that allows to record and process the data from a microphone, 12 body-worn 3D acceleration sensors, and a location estimation. We have evaluated the approach on three different data sets, with a maximal length of two days. 1

Sensor-based statistical models promise to support a variety of advances in human-computer interaction, but building applications that use them is currently difficult and potential advances go unexplored. We present Subtle, a toolkit that removes some of the obstacles to developing and deploying applications using sensor-based statistical models of human situations. Subtle provides an appropriate and extensible sensing library, continuous learning of personalized models, fully-automated high-level feature generation, and support for using learned models in deployed applications. By removing obstacles to developing and deploying sensor-based statistical models, Subtle makes it easier to explore the design space surrounding sensor-based statistical models of human situations. Subtle thus helps to move the focus of human-computer interaction research onto applications and datasets, instead of the difficulties of developing and deploying sensor-based statistical models. Author Keywords Toolkits, Subtle, sensor-based statistical models, machine

Applications that use sensor-based estimates face a fundamental tradeoff between true positives and false positives when examining the reliability of these estimates, one that is inadequately described by the straightforward notion of accuracy. To address this tradeoff, this paper examines the use of Receiver Operating Characteristic (ROC) curve analysis, a method that has a long history but is under-appreciated in the human computer interaction research community. We present the fundamentals of ROC analysis, the use of the A ' statistic to compute the area under an ROC curve, and the equivalence of A ' to the Wilcoxon statistic. We then present several case studies, framed in the context of our work on human interruptibility, demonstrating how ROC analysis can yield better results than analyses based on accuracy. These case studies compare sensor-based estimates with human performance, optimize a feature selection process for the area under the ROC curve, and examine end-user selection of a desirable tradeoff.

An important challenge in designing ubiquitous computing experiences is negotiating transitions between explicit and implicit interaction, such as how and when to provide users with notifications. While the paradigm of implicit interaction has important benefits, it is also susceptible to difficulties with hidden modes, unexpected action, and misunderstood intent. To address these issues, this work presents a framework for implicit interaction and applies it to the design of an interactive whiteboard application called Range. Range is a public interactive whiteboard designed to support co-located, ad-hoc meetings. It employs proximity sensing capability to proactively transition between display and authoring modes, to clear space for writing, and to cluster ink strokes. We show how the implicit interaction techniques of user reflection (how systems indicate to users what they perceive or infer), system demonstration (how systems indicate what they are doing), and override (how users can interrupt or stop a proactive system action) can prevent, mitigate, and correct errors in the whiteboard’s proactive behaviors. These techniques can be generalized to improve the designs of a wide array of ubiquitous computing experiences.

Abstract—Context-aware computing describes the situation where a wearable/mobile computer is aware of its user’s state and surroundings and modifies its behavior based on this information. We designed, implemented, and evaluated a wearable system which can learn context-dependent personal preferences by identifying individual user states and observing how the user interacts with the system in these states. This learning occurs online and does not require external supervision. The system relies on techniques from machine learning and statistical analysis. A case study integrates the approach in a context-aware mobile phone. The results indicate that the method is able to create a meaningful user context model while only requiring data from comfortable wearable sensor devices. Index Terms—Location-dependent and sensitive, wearable computers, mobile computing, machine learning, wearable AI, statistical models. 1

In order to develop intelligent systems that attain the trust of their users, it is important to understand how users perceive such systems and develop those perceptions over time. We present an investigation into how users come to understand an intelligent system as they use it in their daily work. During a six-week field study, we interviewed eight office workers regarding the operation of a system that predicted their managers ’ interruptibility, comparing their mental models to the actual system model. Our results show that by the end of the study, participants were able to discount some of their initial misconceptions about what information the system used for reasoning about interruptibility. However, the overarching structures of their mental models stayed relatively stable over the course of the study. Lastly, we found that participants were able to give lay descriptions attributing simple machine learning concepts to the system despite their lack of technical knowledge. Our findings suggest an appropriate level of feedback for user interfaces of intelligent systems, provide a baseline level of complexity for user understanding, and highlight the challenges of making users aware of sensed inputs for such systems. Author Keywords Intelligent systems, context-aware, mental models,