Browse Prior Art Database

Visual Dialog Showing Speech Interaction with an Intelligent Agent

IP.com Disclosure Number: IPCOM000117234D
Original Publication Date: 1996-Jan-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 4 page(s) / 122K

Publishing Venue

IBM

Related People

Cox, N: AUTHOR [+7]

Abstract

Disclosed is a visual dialog window containing a textual record of a conversation occurring between the user of a computing system with speech recognition capability and with an intelligent agent. While the intelligent agent has been created within the computing system to facilitate more natural and familiar types of interaction between the user and the computing system, permitting the use of either constrained or natural spoken language, the use of natural spoken language is an inherently inaccurate form of input to the system. The visual dialog window allows the user to read his response, as it is understood by the agent, and to determine whether the agent understood his request, so that corrections can be made as needed.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Visual Dialog Showing Speech Interaction with an Intelligent Agent

      Disclosed is a visual dialog window containing a textual record
of a conversation occurring between the user of a computing system
with speech recognition capability and with an intelligent agent.
While the intelligent agent has been created within the computing
system to facilitate more natural and familiar types of interaction
between the user and the computing system, permitting the use of
either constrained or natural spoken language, the use of natural
spoken language is an inherently inaccurate form of input to the
system.  The visual dialog window allows the user to read his
response, as it is understood by the agent, and to determine whether
the agent understood his request, so that corrections can be made as
needed.  Thus, the agent window provides a scrolling history of
commands, reference objects, and other communications.

      The Figure is an example of the visual dialog window, which is
divided into three columns.  A left object repository column holds
objects representing retrieved information and objects the user hands
off to the agent for use in a command.  A central place holder or
reference column includes three-digit numbers that the user can use
to refer to previous parts of the dialog.  A right textual echo
column lists the spoken text provided by both the user and the
system, with a color difference being applied to differentiate these
sources of text.

      In the example of the Figure, the user begins in step 001 with
a specific command to the agent, which the user has previously chosen
to identify as "Zodiac."  The system places a call to the
calendar/scheduler and responds in step 002 with a "to do" list
represented by the object in the left column and by the text in the
right column.  The speaker then gives a command to call "Mary," which
the agent misinterprets as "Terry," as shown in steps 003 and 004.
The user then stops this inappropriate action in step 005, and issues
the command again in step 006.

      Later, in step 015, the user asks the system to get an aerial
map.  In step 016, the system indicates that it has found the map,
showing a map object, but does not display it, since the user has not
asked for that yet.  In step 017, the user asks for the map to be
displayed and printed.  The agent opens a window, in which the ma...