nach oben

2008 | Buch

Kapitel lesen Erstes Kapitel lesen

Modeling Communication with Robots and Virtual Humans

Second ZiF Research Group International Workshop on Embodied Communication in Humans and Machines, Bielefeld, Germany, April 5-8, 2006, Revised Selected Papers

herausgegeben von: Ipke Wachsmuth, Günther Knoblich

Verlag: Springer Berlin Heidelberg

Buchreihe : Lecture Notes in Computer Science

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

Embodied agents play an increasingly important role in cognitive interaction technology. The two main types of embodied agents are virtual humans inhabiting simulated environments and humanoid robots inhabiting the real world. So far research on embodied communicative agents has mainly explored their potential for practical applications. However, the design of communicative artificial agents can also be of great heuristic value for the scientific study of communication. It allows researchers to isolate, implement, and test essential properties of inter-agent communications in operational models. Modeling communication with robots and virtual humans thus involves the vision of using communicative machines as research tools. Artificial systems that reproduce certain aspects of natural, multimodal communication help to elucidate the internal mechanisms that give rise to different aspects of communication. In short, constructing embodied agents who are able to communicate may help us to understand the principles of human communication. As a comprehensive theme, “Embodied Communication in Humans and Machines” was taken up by an international research group hosted by Bielefeld University’s Center for Interdisciplinary Research (ZiF – Zentrum für interdisziplinäre Forschung) from October 2005 through September 2006. The overarching goal of this research year was to develop an integrated perspective of embodiment in communication, establishing bridges between lower-level, sensorimotor functions and a range of higher-level, communicative functions involving language and bodily action. The present volume grew out of a workshop that took place during April 5–8, 2006 at the ZiF as a part of the research year on embodied communication.

Inhaltsverzeichnis

Frontmatter

From Annotated Multimodal Corpora to Simulated Human-Like Behaviors

Abstract

Multimodal corpora prove useful at different stages of the development process of embodied conversational agents. Insights into human-human communicative behaviors can be drawn from such corpora. Rules for planning and generating such behavior in agents can be derived from this information. And even the evaluation of human-agent interactions can rely on corpus data from human-human communication. In this paper, we exemplify how corpora can be exploited at the different development steps, starting with the question of how corpora are annotated and on what level of granularity. The corpus data can be used either directly for imitating the human behavior recorded in the corpus or rules can be derived from the data which govern the behavior planning process. Corpora can even play a vital role in the evaluation of agent systems. Several studies are presented that make use of corpora for the evaluation task.

Matthias Rehm, Elisabeth André

Modeling Embodied Feedback with Virtual Humans

Abstract

In natural communication, both speakers and listeners are active most of the time. While a speaker contributes new information, a listener gives feedback by producing unobtrusive (usually short) vocal or non-vocal bodily expressions to indicate whether he/she is able and willing to communicate, perceive, and understand the information, and what emotions and attitudes are triggered by this information. The simulation of feedback behavior for artificial conversational agents poses big challenges such as the concurrent and integrated perception and production of multi-modal and multi-functional expressions. We present an approach on modeling feedback for and with virtual humans, based on an approach to study “embodied feedback” as a special case of a more general theoretical account of embodied communication. A realization of this approach with the virtual human Max is described and results are presented.

Stefan Kopp, Jens Allwood, Karl Grammer, Elisabeth Ahlsen, Thorsten Stocksmeier

The Recognition and Comprehension of Hand Gestures - A Review and Research Agenda

Abstract

In this paper I review current and past approaches towards the use of hand gesture recognition and comprehension in human-computer interaction. I point out properties of natural coverbal gestures in human communication and identify challenges for gesture comprehension systems in three areas. The first challenge is to derive the meaning of a gesture given that its semantics is defined in three semiotic dimensions that have to be addressed differently. A second challenge is the spatial composition of gestures in imagistic spaces. Finally, a third technical challenge is the development of an integrated processing model for speech and gesture.

Timo Sowa

Modeling Facial Expression of Uncertainty in Conversational Animation

Abstract

Building animated conversational agents requires developing a fine-grained analysis of the motions and meanings available to interlocutors in face-to-face conversation and implementing strategies for using these motions and meanings to communicate effectively. In this paper, we sketch our efforts to characterize people’s facial displays of uncertainty in face-to-face conversation. We analyze empirical data from human–human conversation and extend our platform for conversational animation, including RUTH (the Rutgers University Talking Head), to simulate what we find. This methodology leads to a range of new insights into the structure, timing, expressive content and communicative function of facial actions.

Matthew Stone, Insuk Oh

The Evolution of Cognition — From First Order to Second Order Embodiment

Abstract

The capability to behave autonomously is assumed to rely fundamentally on being embedded into the current situation and in the own body. While reactive systems seem sufficient to address these aspects to assure ones surviving in an unpredictable environment, they clearly lack cognitive capabilities as planning ahead: The latter requires internal models which represents the body and the environment and which can be used to mentally simulate behaviours before actually performing one of them. Initially, these models may have evolved in reactive systems to serve specific actions. Cognitive functions may have developed later exploiting the capabilities of these models.

We provide a neuronal network approach for such an internal model that can be used as a forward model, an inverse model and a sensor fusion model. It is integrated into a reactive control scheme of a walking machine, enabling the system to plan its actions by mentally simulating them.

Malte Schilling, Holk Cruse

History and Current Researches on Building a Human Interface for Humanoid Robots

Abstract

In this chapter we overview humanoid robotic research as another thread of ECA research. This chapter consists of two parts: In the first half, we overview the historical research on creating communicative functions for humanoid robots. In these sections, we see how research on realizing a human interface for a humanoid robot has begun and developed, by tracing some very early humanoid robotic research done in 1970. In the later half, we try to survey recent research related to this topic. In these sections, we will have a brief overview of researches, and also, through the process of reviewing, we try to make clear the characteristics of using robots in the ECA research.

Yosuke Matsusaka

Typological and Computational Investigations of Spatial Perspective

Abstract

This paper is part of an ongoing research program to understand the cognitive and functional bases for the origins and evolution of spatial language. Following a cognitive-functional approach, we first investigate the cross-linguistic variety in spatial language, with special attention for spatial perspective. Based on this language-typological data, we hypothesize which cognitive mechanisms are needed to explain this variety and argue for an interdisciplinary approach to test these hypotheses. We then explain how experiments in artificial language evolution can contribute to that and give a concrete example.

Martin Loetzsch, Remi van Trijp, Luc Steels

Modeling Multimodal Communication as a Complex System

Abstract

The overall behavior and nature of complex natural systems is in large part determined by the number and variety of the mechanisms involved – and the complexity of their interactions. Embodied natural communication belongs to this class of systems, encompassing many cognitive mechanisms that interact in highly complex ways, both within and between communicating individuals, constituting a heterogeneous, large, densely-coupled system (HeLD). HeLDs call for finer model granularity than other types of systems, lest we risk them to be not only incomplete but likelyincorrect. Consequently, models of communication must encompass a large subset of the functions and couplings that make up the real system, calling for a powerful methodology for integrating information from multiple fields and for producing runnable models. In this paper I propose such an approach, abstract module hierarchies, that leverages the benefits of modular construction without forcing modularity on the phenomena being modeled.

Kristinn R. Thórisson

Con-tact – On the Problem of the Absence of Eye Contact and Physical Contact in Virtual Interaction

Abstract

The concrete face-to-face interaction can be regarded as a responsive process, characterized by the fact that – beyond our intentions – we always re-act to the implicit or explicit demands of the other. This form of responsivity is essentially determined by the corporal-sensual presence of the communication partner: the expression of the eyes, gestures and body posture, or the rhythm and sound of the voice. These aspects evocate an implicit horizon of meaning which always affects the intentions of the speaker. In responding to the explicit as well as implicit demands of the other, responsibility for the communication process and the other may arise, which is essential for the success and the continuity of the communication process. In this article I discuss whether this kind of responsibility exists in virtual communication when the nonverbal sphere of communication does not exist, and how it might be replaced at least to some extent.

Barbara Becker

True Emotion vs. Social Intentions in Nonverbal Communication: Towards a Synthesis for Embodied Conversational Agents

Abstract

Does a facial expression convey privileged information about a person’s mental state or is it a communicative act, divorced from “true” beliefs, desires and intentions? This question is often cast as a dichotomy between competing theoretical perspectives. Theorists like Ekman argue for the primacy of emotion as a determinant of nonverbal behavior: emotions “leak” and only indirectly serve social ends. In contrast, theorists such as Fridlund argue for the primacy of social ends in determining nonverbal displays. This dichotomy has worked to divide virtual character research. Whereas there have been advances in modeling emotion, this work is often seen as irrelevant to the generation of communicative behavior. In this chapter, I review current findings on the interpersonal function of emotion. I’ll discuss recent developments in Social Appraisal theory as a way to bridge this dichotomy and our attempts to model these functions within the context of embodied conversational agents.

Jonathan Gratch

Facial Deception in Humans and ECAs

Abstract

Deception is a relevant issue for the theories of cognition and social interaction. When we deceive, we influence others through manipulating their beliefs. This paper presents a definition of deception and of its functions in terms of a model of cognition and social action. We define as deceptive any act or omission aimed at making others believe something false or not believe something true about the invironment, our identity or our mental states. A typology of ways to deceive is outlined (omission, concealment, falsification, masking, negation, and false confirmation), and applied to deception in facial and bodily expression. An ECA is presented that can simulate, mask, or suppress facial expressions of emotions. The relationship of deception and politeness is investigated theoretically and through analysis of a video corpus. The results of the analysis are used to determine when an ECA masks, suppresses or simulates emotional expression.

Isabella Poggi, Radoslaw Niewiadomski, Catherine Pelachaud

Theory of Mind as a Theoretical Prerequisite to Model Communication with Virtual Humans

Abstract

Given the weaknesses that most current implementations of conver- sational virtual humans show, it is argued that future developments might benefit from the incorporation of a theory of mind. First, findings concerning the effects of current embodied agents´ communication will be presented. Then, basic principles of human communication are depicted. Drawing on models of perspective taking, common ground, imputing one´s knowledge to others and theory of mind, necessities for the human agent communication are derived. In the next step, recent implementations from various research groups are presented that take first steps into this direction. Finally, conclusions for future research are drawn.

Nicole C. Krämer

Listening Heads

Abstract

In [1] we discussed functions of head movements and gaze. In this paper, we will go into more depth in the classification of various head movements: how they are distinguished in both formal and functional terms. We look at the distribution of a selection of primitive head movements and their related meanings and the way they are composed out of smaller units. This catalogue is not intended to be exhaustive and does not take into account the component items from other modalities such as speech, facial expressions, gestures and posture. The research is motivated by our desire to build systems that can interpret and mimick expressive human behaviour.

Dirk Heylen

Dynamic Field Theory and Embodied Communication

Abstract

Dynamical Field Theory is a neurally based approach to embodied and situated cognition, in which information is represented in continuous activation fields defined over metric spaces. The temporal evolution of activation patterns under the influence of inputs and neuronal interaction is described by a dynamical system, whose stable states, localized peaks of activation, are the units of representation. This approach has been successfully used to capture many elementary forms of cognition. Communication poses the new challenge of understanding how different modalities can be integrated in a continuously unfolding communicative process. In this chapter we give a brief introduction to Dynamical Field Theory in embodied cognition, and discuss extensions of its ideas to embodied communication. We sketch a highly simplified example of how sequence generation may occur in dynamical fields. We apply these concepts to a specific exemplary problem in embodied communication, turn taking, the temporal structure of which we capture in a simple model.

Yulia Sandamirskaya, Gregor Schöner

’I, Max’ - Communicating with an Artificial Agent

Abstract

With the advent of communicating machines in the form of embodied agents the question gets ever more interesting under which circumstances such systems could be attributed some sort of consciousness and self-identity. We are likely to ascribe to an agent with human appearance and conducting reasonable natural language dialog that it has desires, goals, and intentions. Taking the example of ’Max’, a humanoid agent embodied in virtual reality, this contribution examines under which circumstances an artificial agent could be said to have intentional states and perceive others as intentional agents. We will link our examination to the question of how such a system could have selfawareness and how this is grounded in its (virtual) physis and its social context. We shall discuss how Max could be equipped with the capacity to differentiate between his own and a partner’s mental states and under which conditions Max could reasonably speak of himself as ’I’.

Ipke Wachsmuth

Talking to Virtual Humans: Dialogue Models and Methodologies for Embodied Conversational Agents

Abstract

Virtual Humans are artificial characters who look and act like humans, but inhabit a simulated environment. One important aspect of many virtual humans is their communicative dialogue ability. In this paper we outline a methodology for study of dialogue behavior and construction of virtual humans. We also consider three architectures for different types of virtual humans that have been built at the Institute for Creative Technologies.

David Traum

Can’t Get You Out of My Head: A Connectionist Model of Cyclic Rehearsal

Abstract

Humans are able to perform a large variety of periodic activities in different modes, for instance cyclic rehearsal of phone numbers, humming a melody sniplet over and over again. These performances are, to a certain degree, robust against perturbations, and it often suffices to present a new pattern a few times only until it can be “picked up”. From an abstract mathematical perspective, this implies that the brain, as a dynamical system, (1) hosts a very large number of cyclic attractors, such that (2) if the system is driven by external input with a cyclic motif, it can entrain to a closely corresponding attractor in a very short time. This chapter proposes a simple recurrent neural network architecture which displays these dynamical phenomena. The model builds on echo state networks (ESNs), which have recently become popular in machine learning and computational neuroscience.

Herbert Jaeger, Douglas Eck

Backmatter

Titel: Modeling Communication with Robots and Virtual Humans
herausgegeben von: Ipke Wachsmuth
Günther Knoblich
Verlag: Springer Berlin Heidelberg
Electronic ISBN: 978-3-540-79037-2
Print ISBN: 978-3-540-79036-5
DOI: https://doi.org/10.1007/978-3-540-79037-2