|
|
|
|
|
Towards
Cooperative Conversational Robots and Assistants: Grounded
Situation Models - Where words and percepts meet Nikolaos Mavridis Abstract: Towards the
long-term objective of truly cooperative conversational robots and assistants,
an extensible computational model of embodied, situated language agents and
and an implementation of the model in the form of an interactive,
conversational robot are presented. The central idea is to endow agents with
a sensor-updated "structured blackboard" representational structure
called a Grounded Situation Model (GSM), which is closely related to the
cognitive psychology notion of situation models. The GSM serves as a
workspace with contents similar to a "theatrical stage" in the
agent's "mind." The GSM may be filled either with the contents of
the agent's present here-and-now physical situation, or a past situation that
is being recalled, or an imaginary situation that is being described or
planned. Furthermore, the GSM contains descriptions of both physical (such as
objects) as well as mental aspects of situations (such as beliefs of others).
Most importantly, the proposed GSM design enables bidirectional translation
between linguistic descriptions and perceptual data/expectations. An instance
of the model was implemented on "Ripley", a manipulator robot with
touch, vision, and speech synthesis/recognition, which is able to service
manipulation commands, answer questions about the present or past, and even
"imagine" parts of situations through descriptions. Thus, the
effectiveness of the GSM proposal is demonstrated through a real-world
implementation whose abilities are comparable to those implied by a standard and
widely used test of language comprehension for three-year old children (the
Token Test for Children), and in some directions even surpass those
abilities. ______ Nikolaos
Mavridis, PhD, from Host:
Michael Gennert Refreshments will be served. Last modified: Mon Mar 19 19:10:20 EDT 2007 |