Natural, Free Speech Interaction
A traditional speech interface requires a user to remember and use fixed phrases,
to speak grammatically correct language, to exchange a number of single
question and answer dialogues, to patiently listen the system repeats what the user said, etc.
ATOM Spoken Dialogue SDK resolves these restrictions and supports to develop
a system which allows users to speak as naturally and comfortably as
they are talking with human.
Dynamically Replies, Answers, Asks back, Suggests and Takes Actions
No More Fixed Phrases
For example, when you checks your voice message on your cell phone,
you can freely and naturally speak such as "Did I get some phone calls?",
"Did someone call me?", or "Any massage?".
Grammar Development
Multiple Pieces of Information in one Utterance
・
A voice system for air ticket reservation often asks you
many simple questions and you have to answer each of these
which may be "where is your destination?", "When would you like to leave?" and so on.
A system developped using the ATOM Spoken Dialogue SDK understands natural spoken language
so it allows the user's utterance to contain multiple bit of information
(over-answering) to shorten the length of the dialogue.
Now a user can simply say "I would like to leave NY to Tokyo on 24th".
Grammar Development
Asking Questions to Retrieve Insufficient Information
・
When a user's utterance does not contain all necessary information, the system
will determine dynamically what question to ask,taking the dialogue history into account.
Dialogue Development
Link to Database, Intelligently Ask Questions and Suggest
A shopping mall guide application, for example, will ask back a question of
"where can I find shoes?" by saying "Are you looking for men's shoes, ladies' shoes,
Kids shoes or sport shoes?". A communication between the dialogue management system
and the database of shops, items and locations allows this dynamic, smart interaction.
Dialogue Development
Add Character to Your Application
・
You can add character to your application
by determing how, when and what system will speak. This is done through our
proprietary template mechanism.
Contact
Please Contact Us
for more details.
|
