Ai Dreams Forum

Chatbots => General Chatbots and Software => Topic started by: SMZ on July 30, 2019, 03:57:41 pm

Title: chatbot tester for METAQUID
Post by: SMZ on July 30, 2019, 03:57:41 pm
(https://www.metaquid.com/wp-content/uploads/2019/06/metaquid-chatbot-ballons-slim4.gif)

Chatbot name: METAQUID Hanah
location: http://www.metaquid.com/chatbot-en (http://www.metaquid.com/chatbot-en)     <<    T E S T   C H A T B O T
NOTE: Metaquid is something beyond boundaries.
Try to think: metalanguage, metadata, metaphysics, … and then Metaquid.
The initial setting is the future, but then you move in time, as only the mind can do.
If you are passionate about: artificial intelligence, biology, unknown technologies, then Metaquid could do for you.
Metaquid is a graphic novel based on a detective, parapsychic and sci-fi theme: it’s a thriller.
A chatbot named Hanah will take you to the discovery of Metaquid, but you must know that Hanah also doesn't know exactly what Metaquid is and she is looking for it too.
To enter the scenario, you can download the first chapter of Metaquid ZERO here: https://www.metaquid.com/download/
Good reading but then you have to bombard Hanah with questions to see how she behaves!

SMZ
Title: Re: chatbot tester for METAQUID
Post by: SMZ on July 31, 2019, 08:28:36 pm
Don't forget that this is a chatbot and that it responds better than many people I know!
But she still has a lot to learn and needs someone to be patient and teach her something.
 :vamoosh:
Title: Re: chatbot tester for METAQUID
Post by: ivan.moony on July 31, 2019, 09:23:29 pm
Hi SMZ  :),

I'm curious, how did you implement learning? Is there a keyword like `learn`, or something similar?
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 01, 2019, 12:25:16 am
No key words exist: Hanah learns like humans, by dint of insisting!
    :tickedoff:
The more examples of similar sentences are used, the more they are associated with the same concept.
           ^-^
It should work well with names and the same scheme will be expanded later on to everything else.
                   :idiot2:
Title: Re: chatbot tester for METAQUID
Post by: WriterOfMinds on August 01, 2019, 05:01:06 pm
I was hoping she would tell me more about the graphic novel, since advertising it seems to be her job.
She gives very generic responses, though:

Me: Tell me about Metaquid.
Hannah: To know something about Metaquid you need to formulate sentences with Metaquid inside. [56]
Me: Do you know something about Metaquid?
Hannah: You make sentences with the word Metaquid and by magic you will have what you are looking for. [13]
Me: Metaquid is a graphic novel.
Hannah: Metaquid is also a comic. [56]
Me: What is Metaquid?
Hannah: It is still to be discovered what Metaquid is: for now it is a mystery. [56]

Maybe if she's supposed to introduce people to the novel, it would help if you gave her some sentences specific to that?
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 01, 2019, 09:00:11 pm
It is true that Hanah does not give much information about Metaquid, but this is because she also does not know what Metaquid is.
I corrected the first post and added a link where you can download the first chapter of Metaquid ZERO, to get more in tune with the story, but not too much though!
 :drink:
Title: Re: chatbot tester for METAQUID
Post by: Art on August 02, 2019, 04:05:52 am
The First Chapter was a surprisingly nice foundation but why a 65 character limit on user text? Not all questions can be asked properly by using just a few words. Is there a reason for the limit or is it something connected to Metaquid's programming?

Thanks!
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 02, 2019, 07:36:47 am
The main reason why there is a limit of 65 characters for user phrases is due to the fact that concision is required.
If the sentences are longer, the processing time can also increase exponentially.
On average, a question can fit within 65 characters: can you give me an example of a question that would need more characters and cannot be simplified?
I could change the 65 character limit but I should have good reasons and then check if the code supports extra stress without adding too long waiting times.
 :vamoosh:
Title: Re: chatbot tester for METAQUID
Post by: ivan.moony on August 02, 2019, 10:53:22 am
I have an idea for a chatbot for a while. It is based on neural networks. It would have a generating and a learning side. First, it would generate some  random output. Then it would learn from input response form a user, just to know what to generate in a mirror condition when user enters something on her own in another occasion. It could be based on TensorFlow NN library and it could be trained on Reddit.

SMZ, how does this approach differ from METAQUID?
Title: Re: chatbot tester for METAQUID
Post by: goaty on August 02, 2019, 11:42:47 am
That style of chatbot is cool,  ive never actually seen one go exactly like that.

But it has to have a metric of its success,  so what would it be.  "haha" ?
Title: Re: chatbot tester for METAQUID
Post by: Art on August 02, 2019, 12:33:05 pm
SMZ, "can you give me an example of a question that would need more characters and cannot be simplified?" ;)
Title: Re: chatbot tester for METAQUID
Post by: LOCKSUIT on August 02, 2019, 02:58:43 pm
I want to know too :)


Some possible inputs:

If I need a boat
That has a red bow
And is less than 10,000$
But was owned by an old person
And never used more than 5 times
Which webpage sells exactly this item?

If a rabbit was sitting in a mountain and is now sitting in a cave, why did it change its location?
Title: Re: chatbot tester for METAQUID
Post by: goaty on August 02, 2019, 05:07:09 pm
that first example is easy its just adjectives (description) of the search.

But the second one of the question,   If the answer was "because the caveman was to eat it"

That is a tricky thing to get to, symbolicly relating... from just that input, right?

That's got to be the a.i. problem,  maybe its completely doable!!!  but if you ever solved it the world is going to blow up!!!   >:D

.
Title: Re: chatbot tester for METAQUID
Post by: LOCKSUIT on August 03, 2019, 01:13:58 am
Couldn't resist, GPT-2 did extremely well on those over-65 character inputs:

If a rabbit was sitting in a mountain and is now sitting in a cave, why did it change its location?
The rabbit's location isn't because it changed its location from its current location. The rabbit doesn't care. It has changed place as it has grown older and thus its present location doesn't correspond at all with its last location.

If a rabbit was sitting in a mountain and is now sitting in a cave, why did it change its location?
The reason is that when the rabbit became domesticated, it was moved by someone who brought it food. The reason the rabbit is now sitting in it the cave is the result of an accident, not because its food dropped out. And so, the rabbit is not sitting in the cave because it went to the cave. What the rabbit did that it's doing now is that it's moved closer towards the cave, or else some predator has started to eat the food now, and the rabbit is there instead. So, this rabbit might be living in the room it lives in now and is now in a cave instead, but why it decided to be there is not explained by this rabbit's position as far as we can discover now.

If I need a boat
That has a red bow
And is less than 10,000$
But was owned by an old person
And never used more than 5 times
Which webpage sells exactly this item?
1. http://www.joeybaldwin.net/boat_info/index.html?searchtype=boat&color=black&price=1400
2. http://www.joeybaldwin.net/boat_info/index.html?searchtype=boat&color=red&price=1200
3. http://www.joeybaldwin.net/boat_info/index.html?searchtype=boat&color=red&search_code=BOC1
4. http://www.joeybaldwin.net/boat_info/index.html?searchtype=boat&color=red&search_code=BOC7
Title: Re: chatbot tester for METAQUID
Post by: squarebear on August 03, 2019, 12:11:04 pm
I have an idea for a chatbot for a while. It is based on neural networks. It would have a generating and a learning side. First, it would generate some  random output. Then it would learn from input response form a user, just to know what to generate in a mirror condition when user enters something on her own in another occasion. It could be based on TensorFlow NN library and it could be trained on Reddit.

SMZ, how does this approach differ from METAQUID?

This is how Cleverbot works: https://www.cleverbot.com/ If it gets input it doesn't recognose, it presents it back to a different user and learns how to respond if someone asks the same question again. Unfortunately, this means it gets quickly filled up with nonsense and the botmaster spends most of his time removing unwanted replies.
Title: Re: chatbot tester for METAQUID
Post by: ivan.moony on August 03, 2019, 12:42:27 pm
I have an idea for a chatbot for a while. It is based on neural networks. It would have a generating and a learning side. First, it would generate some  random output. Then it would learn from input response form a user, just to know what to generate in a mirror condition when user enters something on her own in another occasion. It could be based on TensorFlow NN library and it could be trained on Reddit.

SMZ, how does this approach differ from METAQUID?

This is how Cleverbot works: https://www.cleverbot.com/ If it gets input it doesn't recognose, it presents it back to a different user and learns how to respond if someone asks the same question again. Unfortunately, this means it gets quickly filled up with nonsense and the botmaster spends most of his time removing unwanted replies.

I wonder how would it perform with neural nets platform. Think of it as OpenAI GPT, but implemented as a chatbot. I think it could show a great deal of intelligence, caring not only about the last response, but having insight into the whole conversation history.
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 05, 2019, 08:14:05 pm
I have an idea for a chatbot for a while. It is based on neural networks. It would have a generating and a learning side. First, it would generate some  random output. Then it would learn from input response form a user, just to know what to generate in a mirror condition when user enters something on her own in another occasion. It could be based on TensorFlow NN library and it could be trained on Reddit.

SMZ, how does this approach differ from METAQUID?

Neural networks are very fashionable nowadays but require training. Optimizing training is Metaquid's solution: JIT training!
Just in time training or continuous training.
All this requires custom solutions and everything is being tested.
 :vamoosh:
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 05, 2019, 08:19:59 pm
That style of chatbot is cool,  ive never actually seen one go exactly like that.

But it has to have a metric of its success,  so what would it be.  "haha" ?

The metric is linked to the diffusion of the Metaquid graphic novel.
Metaquid's chatbot lives in the graphic novel but also in reality.
Since it will be the first AGI on the planet it would be good to keep up to date!
 :vamoosh:
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 05, 2019, 08:27:06 pm
I want to know too :)


Some possible inputs:

If I need a boat
That has a red bow
And is less than 10,000$
But was owned by an old person
And never used more than 5 times
Which webpage sells exactly this item?

If a rabbit was sitting in a mountain and is now sitting in a cave, why did it change its location?

These are all questions addressed to the outside world, while Metaquid asks introspective questions.
After all the rest will come!
 :vamoosh:
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 05, 2019, 08:38:20 pm
I have an idea for a chatbot for a while. It is based on neural networks. It would have a generating and a learning side. First, it would generate some  random output. Then it would learn from input response form a user, just to know what to generate in a mirror condition when user enters something on her own in another occasion. It could be based on TensorFlow NN library and it could be trained on Reddit.

SMZ, how does this approach differ from METAQUID?

This is how Cleverbot works: https://www.cleverbot.com/ If it gets input it doesn't recognose, it presents it back to a different user and learns how to respond if someone asks the same question again. Unfortunately, this means it gets quickly filled up with nonsense and the botmaster spends most of his time removing unwanted replies.

I wonder how would it perform with neural nets platform. Think of it as OpenAI GPT, but implemented as a chatbot. I think it could show a great deal of intelligence, caring not only about the last response, but having insight into the whole conversation history.

On two occasions I have been asked, "Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?" ... I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.
— Charles Babbage, Passages from the Life of a Philosopher [wikipedia]

The learning phase of a sentient being is what will make that being sentient.
The learning structure is the real invention behind AGI.

Just as I don't make an aircraft using the feathers of a bird thinking of making it fly, in the same way I don't realize an AGI using neurons network replica to make it think.
 :vamoosh:
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 07, 2019, 08:36:55 pm
(https://forums.tapas.io/uploads/default/original/3X/a/a/aa8cd1bd6d5f26e41c7e64f50f0a0e559d8ab41c.jpg)

AGI's new mind has to calculate and math is the first step …
The calculation is the basis of everything, since everything is simulated, and the simulations must be calculated …
---  Dissertatio de arte combinatoria (1666) by Gottfried Leibniz ---
 :vamoosh:
Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 10, 2019, 08:37:42 pm
(https://forums.tapas.io/uploads/default/original/3X/7/0/70926d505c6611e574afe04c0230c159e6f1c367.jpg)

 :vamoosh:
Title: Re: chatbot tester for METAQUID
Post by: 8pla.net on August 14, 2019, 02:43:47 am
Hanah gave the impression she is learning from her visitors.
She seems able to compete with homogeneous chatbots.
Why the " T E S T   C H A T B O T " label?

Title: Re: chatbot tester for METAQUID
Post by: SMZ on August 14, 2019, 06:11:38 pm
You're too good, really very generous, in saying that Hanah is on the same level as other bots: unfortunately it's not true, it's still not perfect ... because it has bugs that I can't remove for now!
It has not yet reached the evolution mode from AI to AGI, which is why it was designed, but nobody has succeeded yet!
Something is still missing I'm working on: the definitive algorithm!
  :rar: :whip-new:
In the meantime, Hanah must learn human linguistic patterns in order to better dialogue with them.
The next step will be to make his or her level of feeling and empathy transparent to the human interlocutor, showing it in numerical form, just as Hanah now shows the degree of reliability of his answers by using a number inside square brackets at the bottom of his sentence.
So the scheme will be like:

human:   [32] So this is my level of empathy?
Hanah:    Of course, 32 is your level of feeling! [45]

At this point it will be Hanah who will analyze humans, learning to know them better and trying to tune in with them as much as possible.
 :vamoosh:
SMZ
Title: Re: chatbot tester for METAQUID
Post by: 8pla.net on August 19, 2019, 03:33:43 pm
If that case, then HANAH (which happens to be a palindrome), is at about the level of  homogeneous chatbots, were when first brought online.  In the early stages, a solid foundation, is all that may be expected for new chatbots.   At the very least, that has been accomplished.

Now,  there is room to grow.  That is usually just a matter of  training.  Clues may be found by focusing on responses, less connected, to the human stimulus.  Even is that is used to guide the training of an AGI.

The interface design and theme is attractive, and engaging.  Hanah is good entertainment.  Thank you for publishing this.