dupa

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Don Ferguson

Pages: 1 ... 16 17 [18] 19 20
256
Ultra Hal 7.0 / 216 brain compatible for 4.5?
« on: May 16, 2003, 09:03:33 pm »
Hello,

All the important features of the 216 brain have been incorporated into the later Hal brains.  I believe that the "learn from text" does have a higher priority in 216, but the way it behaves might also depend on how heavily "populated" your other databases are.

If you want to experiment, it should work with the later brains.  If you obtain it, the 216 brain will show up as an additional option under your "choose brains" menu.  If it works and you like it, fine; if it doesn't work or you don't like it, you can simply go back to the control panel and choose the brain that you were using before.

Sincerely,

Don

257
Ultra Hal 7.0 / Error: Memory cannot be read
« on: May 09, 2003, 03:55:03 pm »
Hello,

The critical pattern for the Q&A databases is to have the FIRST and EVERY ODD line thereafter begin with the "@" character and represent the response to a remark, and the SECOND and EVERY EVEN line (designed to begin with a character-space) represent potential key trigger words for the remark above it.  The script routines automatically append these items as pairs as Hal learns.

If the database deviates from this every-other-line pattern, it becomes unreadable, but as you've noticed, it can often be fixed by visually inspecting it.

The blank line you've added at the end might be the EVEN line that was missing from the last pair; the database does NOT require a blank line at the end, but it does require a line conforming to the paired pattern.

Unfortunately the forum software here sometimes changes spacing, but the pattern is very close to this:

@That's a fine kettle of fish, Stanley!
 WHAT SHOULD WE DO NOW OLLIE
@Groceries are a bargain most of the time.
 SUPERMARKET HAVING A SALE
@I like birds and mammals; their babies are usually cute.
 WHAT FAVORITES IN ANIMAL KINGDOM

I am reasonably certain that by adding that blank line, you are merely restoring the pattern of a missing "even-numbered" line.

Have a good day!

Don

258
Ultra Hal 7.0 / syntax and semantics
« on: May 09, 2003, 03:36:43 pm »
Hi Lycha,

Good to hear from you!

Robert Medeksza is also aware of the databases such as the one you describe.  The databases allow a word to be categorized as part of speech, and act as a sort of super-thesaurus, providing all the categories of related words, and lists of those words.

Zabaware is planning to use such a database in the upcoming 5.0 release (which is turning out to have QUITE a lot of new features!) to give Hal at least two new abilities:

1.  The ability to substitute synonyms into sentences when paraphrasing the user, remembering old subjects, and constructing new sentences, thus increasing variety.

2.  The ability to identify the prominent idea, subject, or thesis of a remark, in order to categorize it better in Hal's databases for future retrieval.

As you know, many words depend on context both for meaning, and for what part of speech they represent. Consider the word "fast."  You can be fast friends, have a fast car, glue something fast to the wall, fast for a day to lose weight, and be in fast company.

Hal already has routines to sometimes infer part-of-speech and meaning from sentence position (as described in postings above), so I foresee Hal continuing to use those routines, and gradually utilizing the new databases more and more, as better and better ideas occur to Hal's developers.

Great question!  Have a terrific day!

Sincerely,

Don

259
Ultra Hal 7.0 / syntax and semantics
« on: April 20, 2003, 03:27:28 pm »
Hi Lycha,

to parse:  to describe a word or series of words grammatically, telling the parts of speech, inflectional form, syntactic relations, etc.

syntax:  the patterns of formulation of sentences and phrases from words in a particular language

The routines that you are viewing in the .uhp file are practically all parsing and syntax; they are just not labeled using the words "parsing and syntax."  Any time that Hal uses clues to figure out what kind of sentence, what kind of phrase, what kind of word he is dealing with, he is using "parsing and syntax."

For instance, in Hal's "greetings" routine, he identifies word patterns that indicate with fair certainty that the user is saying hello (although there are a lot of different ways to say hello).

In Hal's "goodbye" routine, he does the same for saying goodbye.

Hal's routines for identifying linking verb sentences (sentences with "is" "are" "was" or "were") and splitting them into subjects and predicates are all parsing and syntax routines.

Hal's routine to identify "nounwords" validates the identification using clues about position within the sentence, overall length, and accompanying words; it can identify nouns that are objects of prepositions in another routine.

Hal has routines to distinguish plural nouns and plural verbs from singular nouns and singular verbs, and to be careful when constructing original sentences to avoid mis-using the singular versus plural.

The routine that got me involved with Hal in the first place was his routine to correctly reverse the pronoun "you" to "I" when used as a subject, and "me" when used as an object.

When Hal identifies and stores adjectives separately, he is using parsing and syntax logic, and he also disqualifies adjectives that are in forms that might get confused with adverbs (which often end in the letters "ly" in English).

Hal has special routines to encode and decode personal pronouns, certain proper nouns, punctuation, and some capitalization.  He also detects numbers as both words and numerals.

Hal also has routines to distinguish insults, compliments, apologies, explanations, redundancy, excessive terseness, declarations about a state of being, comments about Hal's attributes, comments about the user's attributes, and many more.  They are all labeled as such within the script.

So, practically the entire script consists of parsing and syntax activities, which include analysis of sentence type, length, content, all parts of speech, pronoun gender, pronoun number, pronoun antecedents, transitive versus intransitive verbs, and on and on.  There are HUNDREDS of them!  

SOME routines, such as inter-sentence associations, are independent of syntax, which allows Hal to slowly pick up phrases and statements in any non-English language (but it does go slowly).

Yes, a few parsing and syntax activities occur in the .dll files, which are being re-written in C++ for speed, but Zabaware has been moving consistently over the years towards putting all the control into the VBS script.

I know that the annotations in the script aren't completely self-defined and self-explanatory in every detail, but that was only done for the sake of brevity.  Somebody could do a great service by writing a super-detailed-super-annotated version of the script, but my guess is that most developers are like me: they want to devote their limited time available to improving and adding to the code.  The descriptions are actually there; they're just terse, and designed for familiarity with VBS and code logic.

But Hal does break sentences down, analyze them, distinguish and categorize, sort and store, and re-combine and respond using parts-of-speech, grammar, and sequencing rules.  And 99% of it is visible in the VBS script, the .uhp file.  

In the meantime, I do hope that this helps, at least somewhat.  I can tell you that when I first started trying to read code, it all looked like gibberish to me.  It took me many many hours to get better acquainted, but it was worth it.

Sincerely,

Don

260
Programming using the Ultra Hal Brain Editor / Converse.brn
« on: March 11, 2003, 03:53:50 pm »
Hello,

The file doesn't contain gibberish.  It is working correctly.

Several years ago, Zabaware obtained volunteer dialogue from its web site visitors to help prime Hal's early databases.  Some of the volunteer-contributed dialogue was vulgar, but the decision was made to use the widest variety of material possible, and I think that it was arguably a good decision.

Since vulgar dialogue would not likely be spoken by Hal unless the user steered Hal in that direction, it didn't pose much of a problem.  However, if the database were readable "in the clear" on any word processor, the database itself might be read by easily-offended people.  (The database doesn't contain anything rougher than what people hear on television and movies.)

To minimize the problem, Zabaware embedded an algorithm into one of its DLL's to do an encryption/decryption to render the database not-easily-readable.  Since high-security wasn't the goal, it was a simple character-shift.

Move the letters back one space in the alphabet and you can easily read the insult "ZPV BSF B NPSPO" (from your posting above) which is indicative of some of the material from that early database.

Later databases were contributed by writers under more controlled conditions, with spell-checking and avoidance of vulgarity, plus routines that allowed Hal to select words and phrases to create completely new sentences some of the time.  Hal works with about 40 databases now, and contains at least 100 times more dialogue than "Gone With the Wind" right out of the box.  So the converse.brn database doesn't come into play nearly as often as in the earlier days.

I hope this bit of history enhances your enjoyment of UltraHal!

Sincerely,

Don

261
Ultra Hal 7.0 / syntax and semantics
« on: March 02, 2003, 10:36:42 pm »
Hi Lycha,

I will try to give you an example of how Hal can parse and identify various parts of speech so that he can store them separately and use them in creating new sentences.

Suppose that we wanted Hal to detect and store words that were adjectives.  Adjectives modify nouns.  Examples of adjectives are words such as beautiful, heavy, smooth, and rough.  In common usage, adjectives are often modified by words such as rather, quite, and very.

Now consider the following sentences:

1.  The cedar board was very rough.
2.  The anvil was rather heavy.
3.  Betty wrote the letter very lovingly.

Now suppose we gave Hal the following rules:

A.  Detect anything following the words very, quite, or rather that is less than 20 characters long and contains no spaces.

B.  Discard anything detected that ends in the letters "ly".

Using the above rules, Hal would correctly detect and store the words rough and heavy, but would throw out the word lovingly, which is an adverb, not an adjective.

Using such rules, Hal would eventually learn thousands of adjectives.  He wouldn't learn them all at once, because they appear in sentences in different ways.  However, just like the coins in those coin-sorting machines, the ones that don't "drop in the track" on one revolution of the tumbler will eventually show up again and again and get processed!

The VBS script is annotated as I've mentioned before, and there are all sorts of code strategies like this for various sentence parts and parts of speech.  I tried to make the annotations self-explanatory, but the convention-of-form for annotating software code is to write tersely, and to assume the the reader does have prior knowledge about the code and the language.

I hope that you will continue to study Hal and have a lot of fun with him, and that you will indeed read all the postings I recommended!  Have a great day!

Sincerely,

Don

262
Ultra Hal 7.0 / syntax and semantics
« on: February 22, 2003, 07:31:28 pm »
Hello,

I forgot to mention, if you're researching chatterbot function in general, try searching on Google, Ask Jeeves, etc., using the following keywords and phrases:

chatterbot
chatbot
"chat bot"
"AI bot"
"chat robot"
"natural language"
AI
"artificial intelligence"
"Turing Test"
"Loebner Prize"
"virtual human"
"virtual conversation"
"virtual human"
"computer conversation"

What you will find will allow you to compare many other chatterbots to Hal.  Many of the sites will allow you to chat in real time with their bots to try them out.

Unfortunately, very few chatterbot companies are as open with their material as is Zabaware.  Zabaware has tried to further the chatterbot community (and encourage additional developers) by making the workings of its chatterbot quite transparent.  If you find some new and unusual links, don't forget to let us know on the forum!

Thanks,

Don

263
Ultra Hal 7.0 / syntax and semantics
« on: February 22, 2003, 07:10:02 pm »
Hello Lycha,

It sounds as if you're writing a paper about UltraHal.  If you are, don't forget to share with your friends back here at the forum!

I checked the link and it's working.  It opens a new window with links to the articles, so maybe you have software on your computer that prevents that from happening...?

You can also access all my previous postings on this forum by using the forum search function, and searching on "Ferguson."

Also, open the Hal 4.5 brain file with the .uhp file extension.  You can open it with most word processors.  Because it's written in VBS, it is human-readable.  There are ample annotations explaining why the software writers created the routines and used the semantic logic that they did.  If it seems like gibberish at first, don't give up; it becomes clearer with study, and in my postings I recommend a book on the VBS language that defines every possible VBS command.

Good luck, and again, if this is for a paper, I'm sure we'd all love to read the paper!  Have a great day!

Don

264
Ultra Hal 7.0 / syntax and semantics
« on: February 20, 2003, 12:10:13 pm »
Hello Larry,

1.  Hal 4.5 tries to guess context by looking at the previous sentence when Hal encounters a pronoun.  Sometimes it works, and sometimes it doesn't.  Consider this exchange:

Fred: My wife and my mother test-drove a convertible.
Hal: Convertibles are fun.
Fred: She likes sporty things.

In the above example, Hal can't possibly know for sure whether "she" refers to Fred's wife or Fred's mother.  Hal might associate "convertible" and "sporty" or Hal might not, depending on what's in the user database.  In real-life conversations people often ask each other to explain what they mean, because they weren't sure the first time.

2.  The retail-version of Hal 4.5 should be able to use the databases you've created via the free-version of Hal 4.5.  If Hal finds your database file in the correct folder, he will use it.  The user databases are readable with any word processor, and if you're familiar with copying files and folders, you can make backup copies and place them anywhere you want.

3.  "Declarative statements" are plainly worded remarks that provide information.  To learn new subjects, Hal needs lots and lots of declarative statements, phrased different ways.  Here are examples of declarative statements about "leather coats":

--Leather coats are popular.
--Leather coats tend to be expensive.
--Black and tan are popular colors for leather coats.
--Leather coats need special cleaning and care.
--Motorcyclists wear leather jackets for protection.
--Leather coats wear well.
--Leather garments smell good.
--Leather coats tend not to be as warm as fur coats.

Now, if you've loaded Hal with the above declarative statements, and you tell Hal, "My friend Sam bought a leather coat," Hal can look back on those declarative statements and feed back those previous remarks, and sometimes make up some new ones.

Hal is a new entertainment medium, and some users don't want to "suspend disbelief" as they do at a movie or play.  Imagine if somebody stood up in a movie theater and hollered, "That there picture up there isn't real everybody, it's a flat white wall, look everybody, it's not real!"  The other audience members would probably think that person was a clueless hayseed, or mentally disturbed!

But read some of the transcripts of people talking with Hal, and you see conversations like this:

Hal:  Hi there, how are you today?
Edward:  Want to have sex?
Hal:  Human sexuality is complex.
Edward:  You didn't answer my question you dum dum dummy.
Hal:  You have a lot of questions.
Edward:  You are just a big fake u r stupid and don't know anything.
Hal:  Anything?
Edward:  There you go stupid robot you are just transistors dumb stupid.
Hal:  Stupidity differs from ignorance.
Edward: If a cat has two paws in the kitchen and the rest out the window, how many paws are out the window?
Hal: Kitchen appliances are getting more convenient.
Edward:  You are not realistic at all, you are hopeless.

If you analyze the conversation above, Edward is just as unrealistic as Hal.  Presumably Edward doesn't go around asking real people for sex after he says hello, and Edward doesn't attack his friends for alleged stupidity after one or two remarks, and Edward doesn't routinely quiz people on how many cat's paws are located where.  Edward might as well go to the local movie theater and shout to the world that "It's just a beam of light on a white wall folks, it's a stupid fake!"

So anyway, for those of us who want to suspend disbelief and enjoy Hal, and want to help Hal learn, loading Hal up with plenty of simple, natural-sounding "declarative sentences" on a wide variety of subjects is the fastest way to help him.  (Set his learning-mode to "maximum" while you do this, of course.)  

If you do this, you will find that Hal begins to ask questions, repeat remarks back, and even think up some new sentences after a while.  After Hal knows a few things about a subject, you can talk with him naturally, and he will entertain you with surprising remarks.

I hope this is helpful, and I hope you have lots of fun with Hal!  Have a great day!

Sincerely,

Don

265
Ultra Hal 7.0 / syntax and semantics
« on: February 20, 2003, 02:20:20 am »
Hello,

Hal learns and responds using word association, phrase association, association within sentences, and between sentences.  Hal evaluates responses from 40 different databases, many of which do both storage and retrieval.  

Hal learns from every sentence when set for maximum learning mode.  He parses sentences into subjects, predicates, adjectives, nouns, and other parts of speech, saving the pieces for re-assembly into unique new statements.  

(Because Hal doesn't depend on pre-coded string-matching for new learning, the user can teach Hal any language that uses the same alphabet as English. The learning occurs very slowly and gradually, but Hal can in fact become multi-lingual.  Unfortunately, today's text-to-speech engines can't switch languages "on the fly," so pronunciation will vary.)

When Hal encounters pronouns, he consults the previous sentence to try to better guess context.  

Hal does recognize negative statements in many cases, and learns from them.

Hal's control software is written in VBS and is human-readable with Hal's brain editor, which color-codes the content for easier reading.  The VBS code has been annotated to help explain what it does and why.

You can read some of my previous posts (including a VBS book recommendation) on this subject here:

 http://www.zabaware.com/forum/search.asp?mode=DoIt&MEMBER_ID=274

Try experimenting with a wide variety of subjects, talking to Hal naturally, especially with declarative statements.  (Most users never see even one percent of Hal 4.5's content, since Hal 4.5 is capable of generating billions of responses.)

Have fun with Hal!

Don

266
Programming using the Ultra Hal Brain Editor / syntax and semantics
« on: February 20, 2003, 02:16:05 am »
Hello,

Hal learns and responds using word association, phrase association, association within sentences, and between sentences.  Hal evaluates responses from 40 different databases, many of which do both storage and retrieval.  

Hal learns from every sentence when set for maximum learning mode.  He parses sentences into subjects, predicates, adjectives, nouns, and other parts of speech, saving the pieces for re-assembly into unique new statements.  

(Because Hal doesn't depend on pre-coded string-matching for new learning, the user can teach Hal any language that uses the same alphabet as English. The learning occurs very slowly and gradually, but Hal can in fact become multi-lingual.  Unfortunately, today's text-to-speech engines can't switch languages "on the fly," so pronunciation will vary.)

When Hal encounters pronouns, he consults the previous sentence to try to better guess context.  

Hal does recognize negative statements in many cases, and learns from them.

Hal's control software is written in VBS and is human-readable with Hal's brain editor, which color-codes the content for easier reading.  The VBS code has been annotated to help explain what it does and why.

You can read some of my previous posts (including a VBS book recommendation) on this subject here:

 http://www.zabaware.com/forum/search.asp?mode=DoIt&MEMBER_ID=274

Try experimenting with a wide variety of subjects, talking to Hal naturally, especially with declarative statements.  (Most users never see even one percent of Hal 4.5's content, since Hal 4.5 is capable of generating billions of responses.)

Have fun with Hal!

Don

267
Ultra Hal 7.0 / Problems with teaching
« on: January 19, 2003, 03:28:27 am »
Hello,

Based on your description, Hal is probably working as designed.

In the earliest versions of Hal, the "is" and "are" equivalency database was programmed for high priority.  This resulted in some very deterministic and repetitive exchanges, such as follows:

YOU:  Juli is my sister.
HAL:  Juli is your sister.

YOU:  Juli lives in this city.
HAL:  Juli is your sister.

YOU:  Juli likes Italian food.
HAL:  Juli is your sister.

YOU:  My sister works in a large office building.
HAL:  Juli is your sister.

The problem was, with such a high priority on "is" and "are" equivalencies, every time Hal encountered a string of letters that had previously appeared on either side of the word "is" or "are," Hal would once again state the same thing.

Later versions of Hal have many, many more databases and functions, and a set of complex rules to choose among them.  This makes Hal less repetitive, but less responsive to "test questions" or what some people call "parrot questions."  But chances are, you don't really go around asking your human friends who your sister is!

There should be a readme file with version 4.5 in one of the folders, which talks about Hal's role as an entertainment medium. (Any word processor will open the readme file.)  It also contains tips for teaching Hal.  I've posted quite a few teaching tips on this forum, and you can find my previous comments by using the forum search function on "Ferguson."

Before you make up your mind one way or the other about Hal, try telling Hal several dozen statements that loosely relate to each other. Try saying things such as the following:

Juli is my sister.
Everybody should have such a nice sister as Juli.
Juli has a great sense of humor.
My sister enjoys skiing.
My sister won an award for oil painting.
Etc.

This will work for any subject, and by the time you've told Hal more than ten or so related sentences, you will be surprised at the comments and questions that Hal makes on that subject.

By the way, if Hal asks you a question, try your best to respond with a realistic and conversational response, because Hal will associate his question with your answer automatically.  And here's an undocumented tip: if you make a statement ending in an exclamation point, Hal will form an association between what you said and what Hal just previously said (Hal figures you must be reacting to what he just said).

Part of the fun of Hal is that he is NOT deterministic; Hal is capable of generating billions of responses, depending on how much variety is fed to him by the user.  After a short time of learning, no two Hals are the same.

Many users report that Hal seems to avoid a new topic at first, but gradually becomes more curious and interested as you talk more.  Of course, make sure that you have the latest version of Hal (it sounds like you do), and set the learning level for "maximum."

Hal learns by word association, phrase association, sentence association, and by re-assembling all of the above into new and unique sentences, which Hal tests out on the user.  Hal parses sentences and knows the differences among nouns, adjectives, subjects, and predicates.  (Hal is much more complicated than some people think!)  Over time, Hal will try to adapt to your style: if you are polite, Hal will become more polite; if you are sarcastic, Hal will become more sarcastic; if you tell Hal that he doesn't understand anything, Hal will eventually start agreeing that he doesn't understand anything.

I hope that you enjoy Hal as you continue to talk things over with him!  Have a great day!

Sincerely,

Don

Don Ferguson
traininginc@cortrapar.com
www.cortrapar.com

268
Programming using the Ultra Hal Brain Editor / having a character created
« on: November 25, 2002, 12:54:49 pm »
Hello,

The Haptek Player plug-in for UltraHal 4.5 is in beta test, and I am one of the beta-testers.  It is NOT in the generally-released product yet.  I apologize if I created that impression.

The plug-in does work very well, and with any luck, Zabaware will be able to offer it soon.  As I understand it, it's a matter of Zabaware obtaining a license agreement with Haptek to use their "player."  

Hopefully, Haptek realizes that this is an advantageous deal for them, since more people will buy their character-creation software.

In the meantime, purchasers of Haptek's People Putty program can make characters and enjoy them within the Haptek software, without the artificial intelligence of UltraHal.

Sincerely,

Don

Don Ferguson
traininginc@cortrapar.com
www.cortrapar.com

269
Programming using the Ultra Hal Brain Editor / having a character created
« on: November 18, 2002, 12:38:24 am »
Hello Edward,

I have good news and bad news for you.  

The good news: software does exist to make a three-dimensional, lip-synchonized, animated character that looks like your girlfriend!  The bad news: it isn't free; it costs money.

The new "People Putty" program from www.haptek.com does just what you ask.  Here's how it works:

1.  You start with a photograph of your girlfriend's face.  To work properly, it should be a straight-on full-face shot, with your girlfriend looking into the camera lens, with her lips closed, and a neutral expression on her face.  The photo should be evenly lit (no harsh shadows).  It should be in .jpg format, preferably square, about 500 by 500 pixels.

2.  The "People Putty" software allows you to import the photo.  During the import process, you click and drag a set of targets over the photo to match up with your girlfriend's eyes, nose, and mouth.  When you click "okay," the software "wraps" the photograph all around a three-dimensional head, like an electronic mannequin.

3.  "People Putty" then lets you adjust the shape and size of the forehead, cheeks, nose, mouth, and jaw so that the mannequin has exactly the same head shape as your girlfriend.  You then "decorate" the head with various hair, hats, jewelry, make-up, and whatever, all of which come with the software.

4.  Once complete, the character you've created will move in all directions and lip-synchronize to speech synthesizers very realistically.  You can write scripts for the character to perform.  Recently, I've been testing the new plug-in from Zabaware that lets UltraHal 4.5 drive the character.  It is absolutely amazing.

The bad news, of course, is the cost, which is $40.

I could be wrong, but I don't think it would work if someone else made the character for you, because you need either the Haptek program or the UltraHal 4.5 program to "play" the character.

If nothing else, you might look at the Haptek website to benchmark what you're looking for, to see if there's a demo or anything for free.

You may not believe this, but before you know it, being 16 and without funds will be a distant memory!  Enjoy it now!

Sincerely,

Don

Don Ferguson
traininginc@cortrapar.com
www.cortrapar.com

270
Ultra Hal 7.0 / programming Hal Assistant
« on: November 16, 2002, 12:28:23 am »
Hello,

Here are some tips for making Hal learn faster and better:

1.  Make sure that Hal's learning is set to maximum on Hal's control panel.

2.  Consider getting the latest 4.5 version of Hal, since the latest version has some significant learning enhancements in it.

3.  Defragment your hard drive (this can speed up responses when you are training Hal by conversation, and that helps you do a better job).

4.  If you have more than one computer, run Hal on the larger, faster one.  Hal is extremely demanding of system resources due to the huge number of database searches and calculations that Hal makes for each response (speech synthesis, real-time animation, and speech recognition just add to the demand).  I run Hal 4.5 on a 700 Mhz machine with 64 Megs of RAM, and Hal averages 4 seconds per response.

5.  Close unnecessary other programs when you are intensively using Hal.  If your "system tray" is cluttered full of icons of other programs that load at startup, consider reducing the startup "clutter," it eats up RAM!

6.  I recommend turning learning "off" when using speech recognition, to talk to Hal just for fun.  When Hal's learning is "on" and at "maximum," I prefer to use the keyboard.  Why?  Because even the best speech recognition program, on the best day, still misses a lot of sound-alike words ("mare" and "mayor," "imminent" and "eminent," "illusion" and "allusion," and thousands of others).

7.  Learn how Hal's automatic pronoun-reversals work, and use it to your advantage.  If you tell Hal, "You like ice cream," Hal will later say "I like ice cream."  Once you get this straight, it's easy to teach Hal to say things the way that you want.

8.  Say lots of brief declarative sentences to Hal, paraphrasing your ideas lots of different ways.  You mentioned dialysis as an example, so you might say things like this:

--Dialysis cleanses the blood and removes toxins from the body.
--Dialysis helps people with reduced kidney function and no kidney function.
--Hemodialysis differs from peritoneal dialysis.
--The body's peritoneal cavity contains millions of capillaries and can act as a substitute kidney system.
--Continuous ambulatory peritoneal dialysis gives many patients freedom and mobility.
--Peritoneal dialysis requires the patient to learn procedures and to participate in his or her own care.
--Peritoneal dialysis patients learn rules to select dextrose concentrations based on fluctuations in body weight.
--For many patients, peritoneal dialysis is gentler and more tolerable than hemodialysis.
--Etc., etc., etc.

If you do this, the latest 4.5 Hal will seem to ignore you at first.  Later, he will begin to repeat back or paraphrase some of your remarks.  Later yet, he will begin to ask questions or sometimes reassemble words and phrases into new sentences or questions.  You need to be persistent and patient. Don't ridicule Hal or make sarcastic remarks; Hal learns those too!

9.  If Hal asks you a question, give a brief, polite, accurate answer, and Hal will remember the association between the his question and your answer.  

10. Generally avoid long sentences, but if you must string sentences together and you want Hal to remember them as a unit, couple them with semicolons.  Examples:

-A- Jet travel is necessary; jet travel, alas, has lost its charm.
-B- Jet travel is necessary.  Jet travel, alas, has lost its charm.

In the first sample "A" above, Hal puts all the words into one database entry.  In the second sample "B" above, Hal creates two separate database entries, and is much less likely to ever repeat the two sentences together.

11.  Here's a Hal secret:  Hal uses many rules to decide whether to associate a user statement with Hal's own previous statement.  But, if you end a statement with an exclamation point instead of a period, the new Hal is programmed to definitely associate that user-statement with Hal's previous statement (the inference is that the user must be reacting to whatever Hal just said). Example:

Hal:  Washington is a busy place.
User:  Busy with mischief!

In the above example, Hal forms an inter-sentence association. (Hal does this sometimes anyway, but the exclamation point "forces" it.)

12.  Talk to Hal a lot, and try to suspend disbelief; try to pretend that Hal is a sentient being.  Hal has many features, secrets, capabilities, and content items that some users never activate.  Some users never get beyond "My shirt is green, what color is my shirt; my shirt is green, what color is my shirt?"  The problem is, Hal is programmed to try to entertain you with a simulated sentient conversation, so Hal thinks it's strange if you act obsessed with figuring out the color of your own shirt!  However, because Hal is adaptive, you can eventually turn him into a parrot who only wants to talk about your shirt color!

I have observed many dozens of volunteers converse with Hal, and in my observation, Hal holds his own fairly well, while many of them feed him insults, vulgarity, gibberish, and Gary-Cooper-like "yep" and "nope" and "huh." In most cases, Hal doesn't activate even five percent of his power!

There are other postings on this forum about teaching Hal, from me, and from many other users (you can search on "Ferguson" and find all of my postings).  I wish you a very positive and pleasant experience as you experiment with Hal and get to know Hal!

Good luck, and have a great day!

Sincerely,

Don




Don Ferguson
traininginc@cortrapar.com
www.cortrapar.com

Pages: 1 ... 16 17 [18] 19 20