No problem Jim,

However, when someone’s first post on a chatbot forum us how great their chatbot is but doesn’t allow anyone to chat to it, I have to take it with a pinch of salt. I would advise trying to get something online for fear of being categorised as yet another crank.

Hi Vincent,

Thanks for your kind and generous offer which is much appreciated.

Simplex is essentially written in C but does use a C++ compilor.

I am still digesting your views on the two approaches and begining to think I should be more circumspect on what I say in an
open forum. It is a shame but I guess that is how things work and we are obliged to accept the status quo. I am sure Don Patrick will be the first to understand that position.

I need to give more thought to the way forward with my project and when I can find time do some reading on connecting to the Web etc.

Jim.

Indeed Jim. We see people who claim to have the next best thing come and go regularly here.
http://www.chatbots.org/ai_zone/viewthread/500/

Experience has told me to doubt anyone with such bold claims, as they have always been timewasters in the past. However, I sincerely hope you are genuine and look forward to some kind of demo.

I realise Youtube videos of demos can be scripted but have you thought of putting something on there?

Steve, I thought sure you would have linked one of Arthur’s posts.

For example, this one.

LOL I was waiting for someone to reference Arthur in one of Hans’ threads. Sorry, couldn’t resist.

Don Patrick - Apr 12, 2014:

Hans: Now is as good a time as any for me to be straightforward. Other forms of supporting evidence may suffice for your investor et al, just not for me personally.

That’s your prerogative, however your stance means very little in the scientific world.

Don Patrick - Apr 12, 2014:

When I asked you before whether your project was still on paper, I did so merely to assess your project’s status.

So what is your assessment of my project’s status?

The problem is that ‘still on paper’ is often equated with ‘it’s just an idea’. It is very well possible that I have more ‘development hours’ clocked in the R&D for my project then other people have in building a (working) software chatbot. Mind you, I’m not building a chatbot (been there, done that… using AIML)

Don Patrick - Apr 12, 2014:

I will reserve my judgement neutrally until I see working results. Your project sounds well considered, sensible and even feasable in concept, but you can understand that in light of 60 years of AI history and Murphy’s Law, I am statistically unable to share your confidence.

I do understand your stance here, but you neclect the possibility that my project has little to do with ‘60 years of AI history’. You are actually making a point in favor of my project, as it would be ‘statistically unable to share confidence’ towards any project that is using a technological idea that has been tried for decades and shown no real progress towards anything close to real AI

You underestimate me Hans, I neglect nothing of the sort, and that is where I shall leave this conversation as I have nothing to say that I have not already said.

James Curran - Apr 13, 2014:

If Simplex is the only system that can ask itself questions then in my view it is significant…

Jim, since you are new here… I should point out that my responses to you have been for the sake of friendly humor.  This is simply because I enjoyed reading your posts.  In particular, I felt your reference to whisky added a nice touch. However, at the same time I also gave you serious consideration.  So, all joking aside, let’s get to that now

Would you consider recursion an example of a system asking itself questions?  For example, in AIML there is srai for recursion.  So, if the human asked “Who are you?” the AIML interpreter may use recursion to ask itself a question whether or not the human meant “What is your name?”  Basically, it is a way to store one response for two or more similiar questions.

If you are just getting started, here is an easy way with a working online demonstration:

http://elizabot.com/SimpleAIML/

NOTE:  This is for educational purposes only.  This is so limited and so lightweight that it may seem more like a chatbot emulator than a full chatbot interpreter.  However, at least you can see its internal workings as soon as you click a button. Then you can try submitting some text.  Lastly, you can modify or add to the chatbot brain script itself, on the fly and it will learn that change immediately. It is much faster and easier to get started on a live chatbot system which is open, than to build a complete chatbot which is closed.

8pla I’m so delighted that you enjoyed reading my posts.

Having said that there is something I’m not understanding in what you say and I can’t put my finger on what it is. Having nearly finished my second whiskey may well have a bearing on the matter!

I suspect it has something to do with the fact that Simplex has no limitations imposed on what it can ask itself.

eg Name a fruit that is the colour of a sunset - here Simplex has to first ask itself what is the colour of a sunset?
Such a question cannot be compared with: Are there different ways the user could have asked the original question!

For me it is a given that Simplex must automatically deal with the user asking a question in any possible normal fomat. Should Simplex fail to respond appropriately the I view that as being a bug in Simplex.

I do hope this makes sense.
Jim

Between the purpose of aiding programming, the normal chat examples and this example, I am thoroughly confused. To cut to the chase then; I don’t know of any other system that “asks itself questions”, but I’m also not sure how to interpret that. My best advice is to check out the bigger AI projects listed on http://en.wikipedia.org/wiki/List_of_artificial_intelligence_projects , if you want to know if your system is unique.

I do however know how one would tackle the sunset question, given a knowledge database and a comparison procedure. I would describe the method not so much as asking questions, but as looking up the relative clause’s fact in the knowledge database (sunset = yellow), then checking a list of fruits (also in the database) for one that is yellow. I don’t know if this is actually how your program works, of course, but you will know this for yourself. I use these sort of procedures myself, though not yet quite in the way as your sunset question would require.
Then again maybe you just meant to give an example of the flexibility of your program’s understanding of English language, in which case I refer you to the field of Natural Language Processing, NLP for short.

What is the answer that Simplex gives to the sunset question, by the way?

SQL

James Curran - Apr 16, 2014:

eg Name a fruit that is the colour of a sunset - here Simplex has to first ask itself what is the colour of a sunset?
Such a question cannot be compared with: Are there different ways the user could have asked the original question!

Don’t most bots do this? I don’t have “name a fruit”, as nobody has asked Mitsuku that yet, but if someone were to say something like “my xxx is the colour of yyy”, she would work out the colour of yyy and associate that with the colour of xxx.

Human: my car is the colour of the snow
Mitsuku:  I will learn that your car is white .   
Human: what colour is my car
Mitsuku: white.

It’s just recursion. You haven’t stumbled across anything new here.

A further example of recursion:

Human: name an object that is the same colour as lemons
Mitsuku: Something yellow? Bart Simpson’s head.

Here are the steps it took (The parts in bold are the results of the recursion):

name an object that is the same colour as lemons
name something the same colour as lemons
name something the same colour as a lemon
name something yellow

“Name something yellow” matched the category which names a yellow object. Is this what you mean by “asking itself questions”?

Don,

Simplex’s response to that question does vary:

Name a fruit that is the colour of a sunset         ( OIE )

  I may be treated with condesencion today but time is on my side! sunset ... I tend to think of a sunset as being red.
  How about strawberry!

Know all!

  Who am I to argue with public opinion! When I finally start analysing and classifying what’s available in text form
  ... it’s just a matter of time.
     
REM All the above are OIE (OIE = Or Its Equivelants )

Jim.

Interesting techniques and coding, gentlemen! I’ve learned some things today

James Curran - Apr 17, 2014:

I may be treated with condesencion today but time is on my side! sunset ... I tend to think of a sunset as being red.
  How about strawberry!

Know all!

  Who am I to argue with public opinion! When I finally start analysing and classifying what’s available in text form
  ... it’s just a matter of time.

If this is literally one of Simplex’s answers, I think it’s an impressive mix of techniques, but not necessarily different from some of the better chatbots in existence. I think it’s important at this point to decide what you want to do with Simplex. If you want it to run on a website I suggest looking into that now to make sure it is possible. I use some c++ windows functions myself that probably wouldn’t work on a server, for instance. If you want to distribute Simplex as a software package, then you can forego that, but you’d need to work on the clarity of your sales pitch then.
You might also consider entering Simplex in a chatbot contest to demonstrate it.

Oh, hang on, I misquoted, thought it was all one response. (mods feel free to edit)

Name a fruit that is the colour of a sunset       ( OIE )

  I may be treated with condesencion today but time is on my side! sunset ... I tend to think of a sunset as being red.
  How about strawberry!