Adaptive Interaction: A Utility Maximization Approach to Understanding Human Interaction with Technology

Abstract

is part of the Synthesis Lecture Series on Human-Centered Informatics. The small volume (under 100 pages) presents a cogent case for why we need theory-driven research in the field of human-computer interaction (HCI). But not just any old theory. The authors argue that theories from the separate disciplines of psychology sociology, management science, and economics by themselves are not sufficient for explaining the burgeoning field of HCI. Instead, an overarching theoretical framework is needed that can be used to predict and explain a diversity of technology interactions, from entertainment to team working. Their theory of adaptive interaction (AI) purports to do this, drawing from a number of classic and lesser known theories in the aforementioned disciplines. Payne and Howes argue that it can provide both researchers and practitioners with a powerful way of understanding and reasoning about why people behave and perform tasks in the ways they do.The AI framework focuses on how we select strategies in our everyday and working lives so as to maximize utility, given varying ecological, social, and cognitive constraints. Hence, whether we are interacting with technology (e.g., a phone, a car, a cockpit, a process control plant, Facebook, or Twitter) or comparing data to make a decision (e.g., diagnosing the presence of an illness from X-rays or deciding which online reviews from hundreds posted to read about a hotel for a city we are visiting), we are all the time maximizing some utility function.But what does this actually mean in the context of HCI and interaction design? Throughout the book, Payne and Howes describe this core concept in terms of the currency of interaction, which we use to make our choices, that involves adapting to the environment and the nature of the task. We may vary our strategy depending on what we think is useful, how information is structured on an interface, and how it is presented or ordered. Sometimes, we use internal resources, and at other times we use or rely more on the context (i.e., what is out there to guide our choices). The choice of strategy depends on the context and the constraints surrounding what we are doing. Whenever we have to decide on something, such as making a response, we will select a strategy (from the possible number available to us) that provides a good or best outcome for us, depending on what is of value to us. This could be the most efficient, the quickest, gives us the most information in the least time, conserves energy, and so on. In a nutshell, we make our choices based on "being better than alternatives" (p. 12), although sometimes they can be suboptimal but still the preferred choice.A nice example of how the theory works in practice and why it has much potential for researchers and designers is how we read and act upon reviews on the Web, such as the millions available on Amazon or Trip Advisor. A designer can present a set of reviews in any number of ways-the questions this raises are which, what, and how? Studying users has shown that they selec...

How do children adapt strategies when drawing on a tablet?

“…Our work seeks to further the scientific understanding of childrens drawing strategies. Our theoretical approach is based on a Bayesian approach to explaining adaptive interaction [14]. Our empirical approach, is based on a series of constrained drawing tasks which incorporate drawing feedback.…”

“…Bayesian decision theory provides one key approach to understanding adaptive interaction [14]. In Bayesian decision theory, the visual-motor strategy relies upon the goal of the movement, the planned duration, the possibility of visual feedback during movement, previous training and intrinsic uncertainty in the motor system [19].…”

“…These differences motivated us to observe how children strategize their drawing movement. To examine this movement, we plan to adopt the adaptive interaction framework into our approach [14]. This framework is designed to tackle the adaptive nature of human behavior.…”

Principles, Techniques and Perspectives on Optimization and HCI

“…It turns out that, while such models are difficult to build -because of the extraordinary flexibility of human cognition -recent developments in cognitive science have the potential to form the basis of a new approach. This approach relies on optimal control theory to derive user behavior from theories of underlying psychological constraints and subjective utility [9]. Optimal design might therefore be facilitated by optimal control theories of the user.…”

Interaction science SIG

“…It gains its power and generality from a mathematical formalization of foraging and from a computational theory, ACT-R, of the mind. Theories of movement planning describe how long it takes to move a pointer to a button and also variation induced by changes in the cost of error [13,15]. Theories of team diagnosis, making use of Signal Detection Theory, show the consequences of communication constraints on the integration of multiple diagnostic decisions [17].…”