Skynet-AI is on-line

  • 25 Replies
  • 13371 Views
*

victorshulist

  • Trusty Member
  • ****
  • Electric Dreamer
  • *
  • 118
Re: Skynet-AI is on-line
« Reply #15 on: February 10, 2012, 11:09:29 pm »
Same with Xerox=copy

lol, yep... or Kleenex = facial tissue... the list goes on!

*

Freddy

  • Administrator
  • **********************
  • Colossus
  • *
  • 6860
  • Mostly Harmless
Re: Skynet-AI is on-line
« Reply #16 on: February 10, 2012, 11:33:39 pm »
You also get :

We met a new person last night but he really showed up my friends.

Which means he made my friends feel bad about themselves or something like that.

Or :

This guy at work really showed me up today...

Meaning much the same thing...picking out one's bad points and so on...trying to make one look bad...

@Merlin; Ogden's English, wow that takes me back a while, I was using that as a starting block for a bot, never finished it.

*

Art

  • At the end of the game, the King and Pawn go into the same box.
  • Trusty Member
  • **********************
  • Colossus
  • *
  • 5865
Re: Skynet-AI is on-line
« Reply #17 on: February 11, 2012, 12:07:10 am »
Soda, Pop, aka "Coke" is also referred to as a "Soft drink" as opposed to one containing alcohol.

One other incident was the phrase, "BOT: You hit a fork in the road, which way do you go?"

To "HIT" anything in the road (here in the USA) indicates a type of impact. The car's tire HIT a
pothole or a curb or a tree alongside or another car.

A famous baseball player, Yogi Berra, had many quips and sayings for which he was credited. One
was, "When you come to a fork in the road, take it!"  ^-^

If my car HIT a fork in the road, I might think that it ran over a dinner ware utensil (spoon, knife...).

It is these little nuances and slangs that tend to give chatbots FITS along with their respective
bot masters! Perhaps one day, we'll get all this!

Just my POV and my $.02
In the world of AI, it's the thought that counts!

*

Merlin

  • Trusty Member
  • **
  • Bumblebee
  • *
  • 46
    • Skynet-AI
Re: Skynet-AI is on-line
« Reply #18 on: February 11, 2012, 02:12:02 am »
@Merlin; Ogden's English, wow that takes me back a while, I was using that as a starting block for a bot, never finished it.

Yes, I started looking at converting things to Ogden's "Basic" but I found it better to go in the other direction. As the bot/parser gets smarter, it converts blocks of words into a single word or concept.

*

Merlin

  • Trusty Member
  • **
  • Bumblebee
  • *
  • 46
    • Skynet-AI
Re: Skynet-AI is on-line
« Reply #19 on: February 11, 2012, 02:16:57 am »
So Victor showed up (identified) some problem patterns.

In the UK we use 'showed up' like the way Merlin posed it.

Hum, interesting.  OK.    That's..... different :)

Another dimension to deal with in chatbot design....  'locality' factor of expressions.

I thought it was a typo.

Is it just "showed up", or sometimes "showed"? Why the "up" ?

Can you guys give me some other typical expressions of this "showed up" construct?

Don't know exactly why the "up". It is similar to:
Are you going to show up?
or
Something just popped up.

Maybe it is from the Jack in the Box where you crank the handle and it pops up.

*

Art

  • At the end of the game, the King and Pawn go into the same box.
  • Trusty Member
  • **********************
  • Colossus
  • *
  • 5865
Re: Skynet-AI is on-line
« Reply #20 on: February 12, 2012, 01:16:16 am »
I guess people are so SLOPPY with their English usage or They weren't taught or are too lazy to apply what they should have learned.

The phrase, "Showed up" means makes or made an appearance as perhaps a guest arriving for a party. "The guest and his date showed up somewhat late for the party." More proper would have been, "The guest and his date arrived somewhat late for the party."

Lots of people append prepositions to the end of sentences like, "Why are you doing that for?" or "Where are you going to?
By simply NOT adding those words "for" and "to" in the above examples, the sentence would have been more properly constructed and would have kept the same meaning.

I've noticed the apparent LACK of the usage of the word "THAT" in UK English where the word WHICH seems to be substituted or used instead. "I am going to move the belt to a larger pulley WHICH will drive the blade wheel faster." We (in the USA would likely use "THAT will drive the blade wheel...." or "I'm using the green stain WHICH shows depth better." We would again use the word THAT in place of WHICH.

There are obviously hundreds of slangs, nuances and expressions that one could employ but just how large and at what cost would doing so be practical? Time, processing speed, country / language specific, etc.?

More mud to wade through. :o
In the world of AI, it's the thought that counts!

*

victorshulist

  • Trusty Member
  • ****
  • Electric Dreamer
  • *
  • 118
Re: Skynet-AI is on-line
« Reply #21 on: February 13, 2012, 01:49:39 pm »
You also get :

We met a new person last night but he really showed up my friends.

Which means he made my friends feel bad about themselves or something like that.

Or :

This guy at work really showed me up today...

Meaning much the same thing...picking out one's bad points and so on...trying to make one look bad...


Oh my God.  Ok, I am a human being, and that makes no sense to me and I wouldn't think in a million years that that was what that meant !!!!

Not living in your part of the world of course is the reason.

My bot will not recognize those semantics.. it, like me, will simply report that it doesn't understand those statements.

This brings up,  "Locality" in semantics.     When my bot goes online, sorry guys, but you will have to add those locality-specific semantics to the system yourselves lol :)

Actually, later I will have the system learn semantics by conversation.

** there seems to be no "global" semantics that fits in all cases **

Perhaps a certain percentage, like 80 or 90 % hopefully, of semantics is common throughout the English speaking world, and 10 to 20% are semantics that only apply to certain parts of the world.

No bot can be blamed for not understanding, universally, all areas, and their semantics,  "out of the box" ... . .no human does.

* a related concept ... and I'm not sure if it only applies to Canada... but we have this thing,  "ay"... where we say  "You're going to the party, ay?"  which means  "You're going to the party,  *right*??"  It literally means "is this correct? or "do you agree"?"  with the sentence right before the "ay".    By the way, if you heard about the "eh" in Canada.... "ay" and "eh" are different.  "eh" in Canada is a French Canadian pronunciation problem (french can't say the English "th".. so they pronounce "that" as "dat")  I'm not French.. Polish/German descent.     But "eh" and "ay" are quite different.   

I'm interested though -- how much work are all of you going to put into accommodating "expressions" for every area of the world?  My bot will support 'ay' , but not 'eh'.  It's a Canadian bot after all :)

Art -- your point is well taken regarding the unnecessary prepositions trailing sentences.   Bots will have to do MUCH work in permutations for this.  They will have to make a first attempt at parsing, oh, bad parse? ok, let's try removing this word, ok, try again,  nope,  try adding this word, try again, ok. . now the parse makes sense.

It will just add a LOT of unnecessary CPU cycles, and longer wait time to respond, by the fact that it has to do all this extra work.

the bot has to --

a) find misspelled words...  generate all possible "closest" matches.. try each one

b) determine if there are any unnecessary words,  try removing each one (or some combination of), try again

c) determine if there are words that *should* have been there, but were not (sometimes laziness causing people to remove 'to' for example in "I want create a chatbot".. missing 'to'.... changes the meaning.

so, there will be one thing to having a bot that can carry a conversation intelligently, assuming half decent input, and another problem altogether to make it handle varying degrees of human laziness....   that last one is going to require immense processing time if the input is VERY BAD.
« Last Edit: February 13, 2012, 02:15:49 pm by victorshulist »

*

Merlin

  • Trusty Member
  • **
  • Bumblebee
  • *
  • 46
    • Skynet-AI
Re: Skynet-AI is on-line
« Reply #22 on: February 13, 2012, 02:13:33 pm »
One of the benefits of having your bot on-line, even before you think it is finished, is to get an input corpus that shows some of these language differences. This is one of the reasons that instead of grammar, I try to capture concepts.

So when Victor showed up the other day with some responses that showed up my bot because his input showed up some problems, the fix was pretty easy. ;)

*

victorshulist

  • Trusty Member
  • ****
  • Electric Dreamer
  • *
  • 118
Re: Skynet-AI is on-line
« Reply #23 on: February 13, 2012, 02:21:58 pm »
One of the benefits of having your bot on-line, even before you think it is finished, is to get an input corpus that shows some of these language differences. This is one of the reasons that instead of grammar, I try to capture concepts.

So when Victor showed up the other day with some responses that showed up my bot because his input showed up some problems, the fix was pretty easy. ;)

ah...but grammar is a concept too.  It helps shape the concepts, something people don't realize in chatbot development.... and using an input corpus in that way does not preclude usage of grammar, in fact, grammar will be even more helpful in that context, because you will also understand the structural differences AND the semantic/concept differences and more effectively do the mapping :)    I actually use both grammar and semantic-associations (I guess what you call 'concepts')....  very simple to update also.

I do agree, 100%, that language understanding is utterly impossible with grammar alone.  . .. 

*alone* being the keyword... but without grammar, it is equally impossible.  It is an essential part.
« Last Edit: February 13, 2012, 03:40:51 pm by victorshulist »

*

Merlin

  • Trusty Member
  • **
  • Bumblebee
  • *
  • 46
    • Skynet-AI
Re: Skynet-AI is on-line
« Reply #24 on: February 13, 2012, 04:28:54 pm »
And I do use grammar fragments to aid understanding. We are both approaching the task from different ends, but we may very well end up in the same middle.

One of the things I am trying to do is to extract the meaning of a phrase with as few words possible from the input.

*

Merlin

  • Trusty Member
  • **
  • Bumblebee
  • *
  • 46
    • Skynet-AI
Re: Skynet-AI is on-line
« Reply #25 on: February 25, 2012, 07:51:45 pm »
Thanks again to all who have helped me test Skynet-AI during the last few weeks. It has helped me fix a number of bugs and add a few new features. The Chatterbox Challenge starts next week, if you have time this weekend, please sign-on again and give me feed-back on how your conversation went.

 


Will LLMs ever learn what is ... is?
by HS (Future of AI)
November 10, 2024, 06:28:10 pm
Who's the AI?
by frankinstien (Future of AI)
November 04, 2024, 05:45:05 am
Project Acuitas
by WriterOfMinds (General Project Discussion)
October 27, 2024, 09:17:10 pm
Ai improving AI
by infurl (AI Programming)
October 19, 2024, 03:43:29 am
Atronach's Eye
by WriterOfMinds (Home Made Robots)
October 13, 2024, 09:52:42 pm
Running local AI models
by spydaz (AI Programming)
October 07, 2024, 09:00:53 am
Hi IM BAA---AAACK!!
by MagnusWootton (Home Made Robots)
September 16, 2024, 09:49:10 pm
Attempting Hydraulics
by MagnusWootton (Home Made Robots)
August 19, 2024, 04:03:23 am
LLaMA2 Meta's chatbot released
by spydaz (AI News )
August 24, 2024, 02:58:36 pm
ollama and llama3
by spydaz (AI News )
August 24, 2024, 02:55:13 pm
AI controlled F-16, for real!
by frankinstien (AI News )
June 15, 2024, 05:40:28 am
Open AI GPT-4o - audio, vision, text combined reasoning
by MikeB (AI News )
May 14, 2024, 05:46:48 am
OpenAI Speech-to-Speech Reasoning Demo
by MikeB (AI News )
March 31, 2024, 01:00:53 pm
Say good-bye to GPUs...
by MikeB (AI News )
March 23, 2024, 09:23:52 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

Users Online

416 Guests, 0 Users

Most Online Today: 571. Most Online Ever: 2369 (November 21, 2020, 04:08:13 pm)

Articles