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

Abstract. We describe the design of privacy controls and feedback mechanisms for contextual IM, an instant messaging service for disclosing contextual information. We tested our designs on IMBuddy, a contextual IM service we developed that discloses contextual information, including interruptibility, location, and the current window in focus (a proxy for the current task). We deployed our initial design of IMBuddy’s privacy mechanisms for two weeks with ten IM users. We then evaluated a redesigned version for four weeks with fifteen users. Our evaluation indicated that users found our group-level rulebased privacy control intuitive and easy to use. Furthermore, the set of feedback mechanisms provided users with a good awareness of what was disclosed.

Context-aware intelligent systems employ implicit inputs, and make decisions based on complex rules and machine learning models that are rarely clear to users. Such lack of system intelligibility can lead to loss of user trust, satisfaction and acceptance of these systems. However, automatically providing explanations about a system‟s decision process can help mitigate this problem. In this paper we present results from a controlled study with over 200 participants in which the effectiveness of different types of explanations was examined. Participants were shown examples of a system‟s operation along with various automatically generated explanations, and then tested on their understanding of the system. We show, for example, that explanations describing why the system behaved a certain way resulted in better understanding and stronger feelings of trust. Explanations describing why the system did not behave a certain way, resulted in lower understanding yet adequate performance. We discuss implications for the use of our findings in real-world context-aware applications.

As machine learning (ML) systems emerge in end-user applications, learning algorithms and classifiers will need to be robust to an increasingly unpredictable operating environment. In many cases, the parameters governing a learning system cannot be optimized for every user scenario, nor can users typically manipulate parameters defined in the space and terminology of ML. Conventional approaches to user-oriented ML systems have typically hidden this complexity from users by automating parameter adjustment. We propose a new paradigm, in which model and algorithm parameters are exposed directly to end-users with intuitive labels, suitable for applications where parameters cannot be automatically optimized or where there is additional motivation – such as creative flexibility – to expose, rather than fix or automatically adapt, learning parameters. In our CHI 2008 paper, we introduced and evaluated MySong, a system that uses a Hidden Markov Model to generate chords to accompany a vocal melody. The present paper formally describes the learning underlying MySong and discusses the mechanisms by which MySong‟s learning parameters are exposed to users, as a case study in making ML systems user-configurable. We discuss the generalizability of this approach, and propose that intuitively exposing ML parameters is a key challenge for the ML and human-computer-interaction communities. 1. Introduction and Related

This paper considers how we may design future mobile awareness systems. Building upon research on social cognition, we suggest the need to take into account what is known about humans ’ interpretational capabilities. We identify design issues from the level of an individual awareness cue to the level of a product concept, systematically exposing the associated solution spaces. Using four real applications as analytical examples, we point out multiple ways in which design can affect the user’s processing of awareness information and thereby yield different outcomes in the use of technology. We conclude by pointing out novel design opportunities that lie in the integration of cues with functionality and content on the mobile phone. Categories and Subject Descriptors H.5.3 [Information interfaces and presentation]: Group and Organization Interfaces – asynchronous interaction, synchronous

Designers are skilled at sketching and prototyping the look of interfaces, but to explore various behaviors (what the interface does in response to input) typically requires programming using Javascript, ActionScript for Flash, or other languages. In our survey of 259 designers, 86 % reported that the behavior is more difficult to prototype than the appearance. Often (78 % of the time), designing the behavior requires collaborating with developers, but 76 % of designers reported that communicating the behavior to developers was more difficult than the appearance. Other results include that annotations such as arrows and paragraphs of text are used on top of sketches and storyboards to explain behaviors, and designers want to explore multiple versions of behaviors, but today’s tools make this difficult. The results provide new ideas for future tools. 1.

The first computer users were all programmers, and the field of Human-Computer Interaction started, in part, with a focus on improving how programming was done. There was a significant amount of work in the 1980’s on this topic, but it mostly died out in the 1990s. Now, there is a resurgence of work on what used to be called the Psychology of Programming, Software Psychology, and the Empirical Studies of Programming. Now, research that combines HCI and software engineering concerns regularly wins awards at both the software engineering and HCI conferences, and although there is no longer a conference devoted solely to this topic, it is a major focus of the popular VL/HCC conference series. In this paper, we argue that new HCI and software engineering methods and tools, along with a new acceptance of the programming community, makes it a propitious time for a renewed focus on this topic. One way to define “programming ” is as the process of transforming a mental plan of desired actions for a computer into a representation that can be understood by the computer [10] Expressed this way, it seems obvious that the study of humans and programming should be a topic of HCI. Indeed, this area of study has a long history, and has many names, including the

In this paper we look at awareness systems that use mobile and ambient devices for collecting or presenting awareness information and operate within an Ambient Intelligence Environment. Our focus is on Pervasive Awareness Systems (PAS) that mediate awareness with the aim to improve the feeling of being connected. In particular, we concentrate on the challenges that are connected to participation in multiple communities, requiring a continuous balancing between the need to “keep in touch ” and to reduce interruptions. In the paper, we discuss software agents as a possible solution and identify the different roles that agents can play in reducing interruptions. 1.

In this paper we look at awareness systems that use mobile and ambient devices for collecting or presenting awareness information and operate within an Ambient Intelligence Environment. Our focus is on Pervasive Awareness Systems (PAS) that mediate awareness with the aim to improve the feeling of being connected. In particular, we concentrate on the challenges that are connected to participation in multiple communities, requiring a continuous balancing between the need to “keep in touch ” and to reduce interruptions. In the paper, we discuss software agents as a possible solution and identify the different roles that agents can play in reducing interruptions. 1.

Interruptions in the workplace are becoming increasingly prevalent due to the proliferation of proactive behavior within communication applications and collaborative practices. Interruptions caused by notifications from communication applications (email, instant messaging clients) or operating systems, phone calls and collocated individuals often cause a forced break in the user's activity as they may require action on the user's behalf or cause them to switch their attention to the incoming request. Research has shown that interruptions at inopportune moments often result in substantial costs to users and their tasks, e.g. frustration and reduced productivity. However, information conveyed by notifications is also often beneficial to users. A current thrust within the HCI community has been to develop solutions that reduce the cost of interruption caused by notifications while maintaining their utility. In this work, we focus on one class of interruption, notifications in the desktop, and present one solution to managing such notifications- intelligently timing their delivery. Our solution is based on a deep theoretical understanding of how humans process information and what moments during a user’s task execution exhibit lower mental