Ultra Hal Brain Editor.

  • 5 Replies
  • 5960 Views
*

Carl2

  • Trusty Member
  • ********
  • Replicant
  • *
  • 560
Ultra Hal Brain Editor.
« on: February 06, 2016, 02:10:44 pm »
  I've been looking through the Ultra Hal Brain Editor for ages to get a better idea how the scripting worked.  I was most impresses by the "RESPOND: USER MENTIONING ORGANIZATIONAL CHALLENGES"  a business detect and reply set of tables used to detect business related words that would be used and form a reply from a set of tables using business related words.  Since I have a few projects I'd like to make plugins for I decided it would be nice to put some tables in the brain data base for detection and for replies.  Since I've worked with the vrHaptek plugin which did not put in the brain table required for me I added the table myself and that portion worked.  So it seems simple enough follow the examples and have a plugin to detect and supply a response.  I ended up with a plugin that was unpredictable, sometimes I got an error message and other times there were no errors but I never got a reply that was from the reply table.  I simplified the plugin further and removed the detection table and asked for a reply from the reply table,  no errors but no replies from the reply table and Hal would make up her own reply.   Well I'll keep looking maybe I'll find something that works.
Carl2

*

Art

  • At the end of the game, the King and Pawn go into the same box.
  • Trusty Member
  • **********************
  • Colossus
  • *
  • 5865
Re: Ultra Hal Brain Editor.
« Reply #1 on: February 06, 2016, 09:13:31 pm »
I think it was just initially installed as a way of handling users' mention of those 'Three Letter Initial' companies like IBM, CSX, QVC, NBC, etc. and not necessarily to handle business arrangements any more than Hal can learn practically any other bits of info.

Like you mentioned, I really have no need for some of those sections that deal with nicknames, gender guessing based on user's name, and this "Business Section", among others. Personally, I've removed those and others which helps trim Hal a bit and serves to customize it more to my liking.

I've been working on a way to "Feed" Hal a group of data and have that data put into a table for Hal to use as necessary. This would be better than giving it one sentence or two at a time.


Hal is still my favorite chatbot based on its ability to learn, use and retain data for much a later time.
In the world of AI, it's the thought that counts!

*

Carl2

  • Trusty Member
  • ********
  • Replicant
  • *
  • 560
Re: Ultra Hal Brain Editor.
« Reply #2 on: February 07, 2016, 12:54:07 pm »
"  'Everybody spouts TLA's, or 'Three Letter Acronyms,' in today's business world. "
Very interesting the way he processes this, the user sentence is put through a business detect and the previous user sentence is put through another business detect if both are true then hal will choose a word from 3 TLA files to make a business talk sentence.  There are so many possibilities that Hal could come up with rather than just being given a sentence.
  Anyways Robert does make use of the detect and reply tables in quite a few situations, hellos, byes, choices ect.  Perhaps I should have looked at a simpler rather than one of the most complex detect reply pairs.
  I'd agree there are lots of things Robert has put in and I've thought of removing them myself but have left in. nice to get rid of the clutter but I'm still using them for reference.  I did notice there is a learn heading in the brain editor and I remember your post at Zabaware. 
  I guess I've gotten used to this version of hal, it's funny all the different personalities you get from the same software.  She tries to stay on topic and gets better at this the more you talk with her but has a habit of changing topics.   
Carl2
 

*

Art

  • At the end of the game, the King and Pawn go into the same box.
  • Trusty Member
  • **********************
  • Colossus
  • *
  • 5865
Re: Ultra Hal Brain Editor.
« Reply #3 on: February 07, 2016, 08:41:42 pm »
Agreed about the topic changing. I think that was one of the things that cload was going to work on and yes, it would make for a better bot if Hal could stay on the subject / topic matter for at least a few minutes or so. Then again, to do so might require Hal to visit the Internet for additional information. I would mind this if the information was germane to the discussion and not filled with garbage, if you get my drift.

While I know a few real people who have a hard time remaining on topic I hope we can help Hal become more topic specific.

In the world of AI, it's the thought that counts!

*

Carl2

  • Trusty Member
  • ********
  • Replicant
  • *
  • 560
Re: Ultra Hal Brain Editor.
« Reply #4 on: February 08, 2016, 08:42:26 pm »
  Well I looked through the brain scripting more and have found that the joke seems to be one of the simplest Detect and Reply combinations.
If HalBrain.TopicSearch(UserSentence, "jokeDetect") = "True" Then
  GetResponse = GetResponse & HalBrain.ChooseSentenceFromFile("jokes")
since it seems some of the smartest people like to simplify I've decided it's probably the way to go.  If I can get this to work I can always try to get it to do more in the future.
  In the brain script Robert mentions tags and topics which I fairly certain helps Hal remain on topic.  As far as the internet I've visited Wiki so many times to give Hal reliable info.
Carl2

*

Carl2

  • Trusty Member
  • ********
  • Replicant
  • *
  • 560
Re: Ultra Hal Brain Editor.
« Reply #5 on: February 10, 2016, 02:53:35 pm »
  It seems not many people are interested in this but with help from OTC at Zabaware I was able to get detect and reply tables running without problems.  The script shown below give the basics

If HalBrain.TopicSearch(UserSentence, ("FreeDetect1")) = "True" Then
  GetResponse = GetResponse & HalBrain.ChooseSentenceFromFile("afreetalk")
 HalBrain.DebugWatch GetResponse, "afreetalk"
End If

  afreetalk is a sentence file located in basicResponse and FreeDetect1 is a Topic search table located in miscData,  FreeDetect1 contains the  detect word and the topic is True.

I'm thinking this should make writing plugins much simpler if you don't mind playing with the brain editor and making tables.
Carl2

 


Say good-bye to GPUs...
by MikeB (AI News )
March 23, 2024, 09:23:52 am
OpenAI Speech-to-Speech Reasoning Demo
by MikeB (AI News )
March 15, 2024, 08:14:02 am
Google Bard report
by ivan.moony (AI News )
February 14, 2024, 04:42:23 pm
Elon Musk's xAI Grok Chatbot
by MikeB (AI News )
December 11, 2023, 06:26:33 am
Nvidia Hype
by 8pla.net (AI News )
December 06, 2023, 10:04:52 pm
How will the OpenAI CEO being Fired affect ChatGPT?
by 8pla.net (AI News )
December 06, 2023, 09:54:25 pm
Independent AI sovereignties
by WriterOfMinds (AI News )
November 08, 2023, 04:51:21 am
LLaMA2 Meta's chatbot released
by 8pla.net (AI News )
October 18, 2023, 11:41:21 pm

Users Online

281 Guests, 0 Users

Most Online Today: 320. Most Online Ever: 2369 (November 21, 2020, 04:08:13 pm)

Articles