|
|
|
|
Brainhat is a computing environment that is programmed in
natural language—particularly English, at present.
Program statements are propositions (facts) and inference
templates (executables).
Taken together, they can motivate goal-oriented
conversations or actions.
Brainhat can interact with you or with other Brainhats.
It can manage outside processes, learn and remember.
It has interfaces for robotics, speech, text and more.
Multiple interfaces can be active simultaneously.
Over the years, we've interfaced with robotics, VoiceXML,
HTML, IM, email and speech recognition for dictation and
telephony.
The project is in the midst of a long development cycle,
with some nice capabilities on the way.
In the meantime, we're keeping the site alive and
providing code updates every once in a while.
|
Here's an illustration of Brainhat
at work. The vocabulary and English grammar are known to the program in advance. The first few statements are
inference templates, followed by some propositions. At the end of the code sample, Brainhat answers "why?" to reveal
the nature of its conclusions.
>> if thing1 is near thing2 then thing2 is near thing1
if thing1 is near thing2 then thing2 is near thing1.
>> if a man sees the princess then he is happy
if a man sees the princess then he is glad.
>> if a man is near the princess then a man can see the
princess
if a man is near the princess then he sees the princess
>> luigi is near the princess
luigi is near the princess. luigi sees the princess. he is glad.
>> why?
he is glad because luigi sees the princess.
>> why?
luigi sees the princess because luigi is near the princess.
|
|
Internally, the captured knowledge is organized as trees, similar to sentence diagrams.
These can be manipulated, stored, recalled and generalized as processing proceeds.
Interestingly, once the tree is created, the words are no longer needed;
the structure stands for captured knowledge in the absense of language.
There quite a bit to Brainhat; we've been working on it for many years.
At some point, the documentation will catch up.
Look around the site.
Experiment if you wish.
I'd love to hear from you.
Kevin Dowd, dowd@atlantic.com
|
|
