People, How can I FIX this horrible thing?

[dihelson]

Hello, my friends,

I don't know, but may be HAL have some problems.
It talk nonsense most of the time. I would like to know how to learn to edit and locate where these kind of expressions are, or what files in order to delete it.

See these terrible examples of conversations:

Me:Do you have dreams?
HAL:What do we look forward to in the near future?


Me:what can you teach me ?
HAL:In their dreams, poodles make wolves cower and whimper.

Me:tell me the name of the planets
HAL:A certain cola has an outstandingly well-known trade-mark.

Me:how can I help you?
HAL:Expectations influence results.


Me:What's the name of our galaxy?
HAL:Cheese has calories, but it tastes so good!

Me:Do you like people?
HAL: Uniforms convey a certain confidence.

It would be interesting that if HAL doesn't want something it simply doesn't say those garbage... It could say what we HUMANS say: I DON'T KNOW. CAN YOU TEACH ME ?
This may be some kind of text file, part of the brain. I would like to find it and delete, or edit. It's better HAL talking less, but correct, instead of saying nonsense.


Thanks in advance.

what's the name of my girlfriend?
Nurses more than earn their money.

Can you teach me?
Fast moving water swirls and gurgles past obstructions.

[James P]

Hi, Some times you need to treach Hal using a "what", "What" or a "Why","Why" meathod. this is what I have done, eg:

ME: Why do birds fly south in the winter?
HAL: Dogs are not cats,

ME: Birds fly south in the winter because........
HAL: Bla Bla Bla Bla

ME: Why do birds fly south in the winter?
HAL: Birds fly south in the winter because........

This sometimes workes but does have its pitfalls as sometimes Hal will answer your question by asking it back to you.

[dihelson]

quote:Originally posted by James P

Hi, Some times you need to treach Hal using a "what", "What" or a "Why","Why" meathod. this is what I have done, eg:

Hello, James,
YES, I discovered this by chance!  
No matter what HAL says, I continue to teach him:

like:
Our Galaxy is called Via-Lactea
Birds can sing

etc
etc...
I have noticed that sometimes it answer to me with the same question I did, or worse: Exchanging I for YOU. like

Me : "Do you like Bach?"
Hal: "You said you like Bach"

Me: I love teaching
HAL: You love Chopin because Chopin is romantic

eheheheh

There's something quite unplesant also:
If I say:

MIKE is a friend of mine
Mike works on a Gas Station

Then I ask HAL:
Who is Mike:
HAL: Mike is your friend
Me: Where does he works ?
HAL: I don't know

I.E - HAL should have been programmed for inside a conversation, when we say HE or SHE, or IT, he KNEW that we are talking about the same Matter of the anterior sentence. If we are talking About MIKE on a sentence, and on next sentence If I use HE instead of repeating MIKE again HAL should suppose I am still talking about MIKE.
I think this is not too hard to implement. We need just keep the last sentence KEYWORD, and on the following sentence, if there's any HE or SHE or IT , HAL could associate with that keyword.

All the best,
Dihelson

[James P]

I dont know if this will help, but what I am doing at the moment (Or when I get back home) is Trolling through the default brain, file by file and deleting all keywords which are in the wronge place or any sentences my bot is saying which usualy end up out of context.

Another thing I have found is that I can not access the Keyword and priority brain through the brain editor when I choose XTF v1.02 in the brain editor (I get a currupt brain error) But if I choose the XTF v1.0 in the brain editor I can edit the Keyword and priority brain and any changes I make, come through when my bot is using the XTF v1.02 brain.

Hope this helps , even a bit

[dihelson]

quote:Originally posted by James P

I dont know if this will help, but what I am doing at the moment (Or when I get back home) is Trolling through the default brain, file by file and deleting all keywords which are in the wronge place or any sentences my bot is saying which usualy end up out of context.

What a coincidence, James, I stayed all morning trying to delete in original brain, or on XFT, those Garbish sentences which generate OUT OF CONTEXT messages. It's hard work. I almost deleted the brain entirelly, Indeed deleting the Question and Answer completelly is what I need, preserving only the detection of Yes and No, gramatics, and these things, but the Database, I'd like to Substitute.


Who like to know about Democrats and Republicans ? There are so much stupid sentences talking about nonsense that it would be a real pleasure to CLEAN this brain of all these things.

I tried another approach this morning also: Building an entire brain from scratch, but although I have talked with him for almost 4 hours, he learnt many things, but it doesn't have LOGIC. It's conclusions are weak, and we have no control about Yes, or No, repetitions, adjetives, substitutions, etc...

If you talk with this brain you feel it is a bit crazy
But I'm reading a lot. I hope I could ELIMINATE ALL RANDOM SENTENCES. That's what I need. Eliminate all Random Sentences. When HAL doesn't know anything it tell: I DON'T KNOW, then I teach him. it's better that listening to horrible things like :"Parrots are green Birds" inside a conversation about Philosophy!

All the best,
Dihelson

[James P]

Unfortunately you will never be able to get rid of all of Hals random responces, I have been using Emma (my bot) now for nearly a year, or thats what it feels like at times, and I still get random comments and out of context statements, i guess it is something we learn to love.

[dihelson]

quote:Originally posted by James P


Another thing I have found is that I can not access the Keyword and priority brain through the brain editor when I choose XTF v1.02 in the brain editor (I get a currupt brain error) But if I choose the XTF v1.0 in the brain editor I can edit the Keyword and priority brain and any changes I make, come through when my bot is using the XTF v1.02 brain.

Hope this helps , even a bit

James,
I think this is because you haven't renamed that
enhanced_main.brn (or something like that) to
XTF_SYS_enhanced_main.brn

You need to do it, in order to access it from the tools interface.

ABOUT THE RANDOM RESPONSES:
I saw on the other UltraHal forum, that some guys are almost finished a brain which has no RANDOM responses. But it is payware.
There are also another called SuperBrain 8 or something like that which is really big in knowledge.

I'd like to know even more links to FREE brains and experiments.

Regards,
Dihelson