Cristen Torrey

Anthropomorphic Interactions with a Robot and Robot–like Agent

People’s physical embodiment and presence increase their salience and importance. We predicted people would anthropomorphize an embodied humanoid robot more than a robot–like agent, and a collocated more than a remote robot. A robot or robot–like agent interviewed participants about their health. Participants were either present with the robot/agent, or interacted remotely with the robot/agent projected life–size on a screen. Participants were more engaged, disclosed less undesirable behavior, and forgot more with the robot versus the agent. They ate less and anthropomorphized most with the collocated robot. Participants interacted socially and attempted conversational grounding with the robot/agent though aware it was a machine. Basic questions remain about how people resolve the ambiguity of interacting with a humanlike nonhuman. By virtue of our shared global fate and similar DNA, we humans increasingly appreciate our similarity to nature’s living things. At the same time, we want machines, animals, and plants to meet our needs. Both impulses perhaps motivate the increasing development of humanlike robots and software agents. In this article, we examine social context moderation of anthropometric interactions between people and humanlike machines. We studied whether an embodied humanlike robot would elicit stronger anthropomorphic interactions than would a software agent, and whether physical presence moderated this effect. At the outset, robots and agents differ from ordinary computer programs in that they have autonomy, interact with the environment, and initiate tasks (Franklin & Graesser, 1996). The marriage of artificial intelligence and computer science has made possible robots and agents with humanlike capabilities, such as lifelike gestures and speech. Typically, “robot” refers to a physically–embodied system whereas “agent” refers to a software system. Examples of humanlike robots are NASA’s Robonaut—a humanoid that can hand tools to an astronaut (robonaut.jsc.nasa.gov/robonaut.html), Honda’s Asimo, and Hiroshi Ishiguro’s

Learning how: the search for craft knowledge on the internet

Communicating the subtleties of a craft technique, like putting a zipper into a garment or throwing a clay pot, can be challenging even when working side by side. Yet How-To content - including text, images, animations, and videos - is available online for a wide variety of crafts. We interviewed people engaged in various crafts to investigate how online resources contributed to their craft practice. We found that participants sought creative inspiration as well as technical clarification online. In this domain, keyword search can be difficult, so supplemental strategies are used. Participants sought information iteratively, because they often needed to enact their knowledge in order to evaluate it. Our description of people learning how allows us to elaborate existing understandings of information-seeking behavior by considering how search originates and is evaluated in knowledge domains involving physical objects and physical processes.

Effects of adaptive robot dialogue on information exchange and social relations

Human-robot interaction could be improved by designing robots that engage in adaptive dialogue with users. An adaptive robot could estimate the information needs of individuals and change its dialogue to suit these needs. We test the value of adaptive robot dialogue by experimentally comparing the effects of adaptation versus no adaptation on information exchange and social relations. In Experiment 1, a robot chef adapted to novices by providing detailed explanations of cooking tools; doing so improved information exchange for novice participants but did not influence experts. Experiment 2 added incentives for speed and accuracy and replicated the results from Experiment 1 with respect to information exchange. When the robots dialogue was adapted for expert knowledge (names of tools rather than explanations), expert participants found the robot to be more effective, more authoritative, and less patronizing. This work suggests adaptation in human-robot interaction has consequences for both task performance and social cohesion. It also suggests that people may be more sensitive to social relations with robots when under task or time pressure.

Evaluating the Effectiveness of a Tutorial Dialogue System for Self-Explanation

Previous research has shown that self-explanation can be supported effectively in an intelligent tutoring system by simple means such as menus. We now focus on the hypothesis that natural language dialogue is an even more effective way to support self-explanation. We have developed the Geometry Explanation Tutor, which helps students to state explanations of their problem-solving steps in their own words. In a classroom study involving 71 advanced students, we found that students who explained problem-solving steps in a dialogue with the tutor did not learn better overall than students who explained by means of a menu, but did learn better to state explanations. Second, examining a subset of 700 student explanations, students who received higher-quality feedback from the system made greater progress in their dialogues and learned more, providing some measure of confidence that progress is a useful intermediate variable to guide further system development. Finally, students who tended to reference specific problem elements in their explanations, rather than state a general problem-solving principle, had lower learning gains than other students. Such explanations may be indicative of an earlier developmental level.

How-To pages: Informal systems of expertise sharing

The How-To has recently emerged as a genre of online content that describes how something is done. This study focuses on computer and electronics hobbyists and their use of How-Tos—how hobbyists use existing knowledge to solve technical challenges, how they document their new knowledge for one another, and how they exchange help and feedback. Our analysis describes How-To knowledge sharing as a fully decentralized expertise-location system in which the How-To functions as both a broadcast of the author’s expertise and a personal portfolio.

How a robot should give advice

With advances in robotics, robots can give advice and help using natural language. The field of HRI, however, has not yet developed a communication strategy for giving advice effectively. Drawing on literature in politeness and informal speech, we propose options for a robots help-giving speech-using hedges or discourse markers, both of which can mitigate the commanding tone implied in direct statements of advice. To test these options, we experimentally compared two help-giving strategies depicted in videos of human and robot helpers. We found that when robot and human helpers used a hedge or discourse markers, they seemed more considerate and likeable, and less controlling. The robot that used discourse markers had even more impact than the human helper. The findings suggest that communication strategies derived from speech used when people help each other in natural settings can be effective for planning the help dialogues of robotic assistants.

Interactivity and expectation: eliciting learning oriented behavior with tutorial dialogue systems

We investigate the reasons behind students’ different responses to human versus machine tutors and explore possible solutions that will motivate students to offer more elaborated responses to computerized tutoring systems, and ultimately behave in a more “learning oriented” manner. We focus upon two sets of variables, one surrounding the students’ perceptions of tutor qualities and the other surrounding the conversational dynamics of the dialogues themselves. We offer recommendations based on our empirical investigations.

CycleTalk: Toward a Dialogue Agent That Guides Design with an Articulate Simulator

We discuss the motivation for a novel style of tutorial dialogue system that emphasizes reflection in a design context. Our current research focuses on the hypothesis that this type of dialogue will lead to better learning than previous tutorial dialogue systems because (1) it motivates students to explain more in order to justify their thinking, and (2) it supports students’ meta-cognitive ability to ask themselves good questions about the design choices they make. We present a preliminary cognitive task analysis of design exploration tasks using CyclePad, an articulate thermodynamics simulator [10]. Using this cognitive task analysis, we analyze data collected in two initial studies of students using CyclePad, one in an unguided manner, and one in a Wizard of Oz scenario. This analysis suggests ways in which tutorial dialogue can be used to assist students in their exploration and encourage a fruitful learning orientation. Finally, we conclude with some system desiderata derived from our analysis as well as plans for further exploration.

Young researchers' views on the current and future state of HRI

This paper presents the results of a panel discussion titled “The Future of HRI,” held during an NSF workshop for graduate students on human-robot interaction in August 2006. The panel divided the workshop into groups tasked with inventing models of the field, and then asked these groups their opinions on the future of the field. In general, the workshop participants shared the belief that HRI can and should be seen as a single scientific discipline, despite the fact that it encompasses a variety of beliefs, methods, and philosophies drawn from several “core” disciplines in traditional areas of study. HRI researchers share many interrelated goals, participants felt, and enhancing the lines of communication between different areas would help speed up progress in the field. Common concerns included the unavailability of common robust platforms, the emphasis on human perception over robot perception, and the paucity of longitudinal real-world studies. The authors point to the current lack of consensus on research paradigms and platforms to argue that the field is not yet in the phase that philosopher Thomas Kuhn would call “normal science,” but believe the field shows signs of approaching that phase.