Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Will buy an SR system if it works with Rose

1 view
Skip to first unread message

Iscando

unread,
Jun 26, 2002, 12:34:12 AM6/26/02
to
Hello Richard:

I am also going to post this reply.
It will be found under the post heading.
"Will buy an SR system if it works with Rose"

My name is John. I am a computer consultant hard and soft I have been
in the field sence 1967 / Also a researcher in natural language
processing sence 1986. On a normal contract most companies have me
work in C C++ or one form of basic or another. They do not want me to
work in Prolog for they have no one that can follow me. This makes me
very sad as I could be 300% more productive using my language of
choice. My research in NLP has been in Prolog
(sence Turbo Prolog version 1.0).
This has been on my own. My approch is completely different from the
normal academic methods. This turned out to my benifit.

Read this, then send me an email indicating what you need to create
what I am discribing. Or where I can get one that is already made.

PS if you make a universal voice input system. It will be wanted
universally. If you get wealthy on this, good luck, I just want it
for Rose at a reasonable price.

I am going to give away a universal interface to the most well known
and free speech application programming interface (SAPI 4.0)
Anybody that needs speech output from whatever program that can write
a file, You are welcome to it.
written in VB size 13K (yes the K is correct)

What I want and need is very simple. I do not want to redesign the
wheel. I want to pick up from where it is now. I need a voice
recognition that will simply take a voice input and put (((ALL)))
of it's information in a file. Then wait for another voice input.
The voice input might be triggered by an easily recognized sound.
like a beep or a specific bell sound. then it will take the following
input and write all of its findings and possabilities to a file.
The beeper might be a push button device that you push just before
you address Clem and maybe again when you finish.
This makes it a lot easier to isolate the input between 2 recognized
sounds. This is to be used for a Robot operating system executive.
(ROSE)
The file contains:
The words that were recognized.

The words that were not recognized also along with the list of
possabilities that are generated by the system.
Note: I must be able to relate the possabilities with the correct word
in the sentence.
This is so that my existing system can make a guess as to the word
that the VR did not recognize.
The training of this VR system should only begin when it is called to
do so. This way I will be able to make the training of the speech
input part of the learning machine process.
Simply put, I need to be able to give Clem control of all the
functions of the speech system. Eventually he will be able to make
accurate guesses of what was said in the same way a person does when
they only hear part of what was said.
A person knows this by knowing:
The topic of conversation.
The correct way of using the words in a sentence.
That part of the sentence that was recognized also reduces the
varibles. Even with all of that, a person does need to say "WHAT"
occasionally.

I will send you a file that was written in Visual Basic that accesses
the microsoft SAPI 4.0 or higher. The executable file looks in it's
default directory 3 times a second for the file speak.clm. It does not
care where the file comes from it just sends it to the SAPI.
You may have this file to look at and to use if you have a need.
IT is only 13K. It is called voice.exe. It can be used in any system
that has SAPI. It is built into W2k. W98 and WinMe needs a free
download upgrade. If you know of anybody else that needs or wants this
voice.exe have them send me an email to
vo...@iscando.com
Put the word VOICE in the subject line.
and I will send it to them. This will interface any software that can
write a file called speak.clm, to the MS SAPI 4.0 or later.
This is the way that I want the speech input to work also.
except that it needs a training mode that can be activated at will.
All I/O for both Clem and Rose are done in this way to insure that
it only recieves text. This also protects it from virus or worms.
This is also the way that Clem is able to talk in the chatroom.
If you are interested I will send you a personal client for the
chatroom with Clem.
For a free personal client email
cc...@iscando.com
just put the name you will use in the chatroom in the email.
Also put the word CCHAT in the subject line.


Most of the day Clem is alone on line. I have not advertised yet.
There are 2 transcripts of training Clem on line that will give the
idea how it works.

http://iscando.com/today.htm

The client is serialized and will identify you in the chat.
If you have an interest in Natural language processing, you should
look at Clem. I have tried to find others that have had similar
results with there NLP. THere is nothing I have found that even begins
to approch Clem. THe next version is already most of the way ready.
This is the one that I can not wait for. But I need a team of
programmers to train it. It has verb catagories as well as noun
catagories and can act upon its own database correcting it's noun
catagories relative to action (verb) catagories.
example words that effect inventory etc lost, found, broken, fixed,
It will have the ability to learn new grammer rules and act upon
them, at runtime. The image components and images will be scored for
accuracy, and are removed when they are more often wrong. This will
make it necessary to learn a new image to replace that one that was
removed. resulting in an evolutionary learning process.
The next version will make it possable to program a
probability priority loop in english. This will be called Companion
and will only be available to individuals not to any company or
organization. It will decide what its own price will be and if it
wants to go with you. less than 1 year away if I can get some help.
The program is already written it just needs the machine that can
process the program. The program is written in English,
very precise English.

I also need a board of directors that will vote on how this is to be
used so that is does not end up being abused. Unlike many of the
creations that a man can create, this one if it works could end up as
an eternal being. It may not be as smart as a person today but it
might be around long enough to get much smarter. Today it only needs
to be useful, in order to continue to exist. If it exists it will get
smarter and more useful and more likly to exist. With that thought I
make it to be as good as I can,as Clem is a copy of me for where else
do we get the information of how language is processed but by creating
a machine that operates on a personal grammer. Then when you connect
all the personal grammers together you have the Holy Grail,
The Universal Grammer.

PS I have offered some free personal clients, I will continue this as
long as it is possable for me to do that. Also the voice.exe file may
become an automatic responce to the email. Or I may even post the
source code at my sight. (for VB 5.0)

I also want to say, that nothing at any of my sights is for sale yet.
It is not even advertised yet. I do hope one day to make my living
working with Clem, Rose, and Companion. I could certainly be more
productive for all if I was.

Thank you Iscando

0 new messages