Ai Dreams Forum

Artificial Intelligence => General AI Discussion => Topic started by: darkwolf77 on January 02, 2007, 02:44:49 am

Title: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 02, 2007, 02:44:49 am
Hi!

I just joined DigitalGirl, but have been reading the posts here for quite some time.  I've been trying out the various AI chatterbots and desktop assistants for awhile now and unfortunately haven't found any that really impress me.  Currently, I'm playing around with KARI 2.6 and doing graphics for her.  You can find some of my graphics in the file sharing topic on KARI's forums.

The project I want to organize would make today's desktop AI's obsolete.  I need experienced programmers willing to donate time and effort to create an open-source AI assistant with a graphical interface (2D for now, 3D later)  I can supply the 2D graphics, and I have a good idea of how I want to implement the project, but not the programming skills to put it into code.

Eventually, I'd like to implement Haptek or perhaps an open-source 3D engine to create a life-like desktop assistant that not only learns, but learns rapidly from simple comversations.  The project would allow the bot to search web resources, such as wikipedia or a custom list of websites for additional data.  This would allow the bot to gain knowledge even faster and make it infinitely customizable.

I want the bot to be able to work with a desktop interface and be aware of the system.  I'd also like to streamline the code to work well on lower-end hardware.

Unlike a lot of folks, I do not believe that the future of AI will be web interfaces, but human-to-machine interfaces that will allow us to communicate with our machines in a much more natural fashion.  I do not want to use MS-Agents, but a proprietary graphic interface similiar to KARI, but with more customizable graphics options

I'll be honest, I'm more of a project leader than a lead programmer.  I left off programming many years ago and turned to the hardware side of computing.

If you're interested, please post here.  I'll be checking the board frequently and hope to hear from some of you soon.

With Regards,
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 03, 2007, 11:55:22 pm
Okay.

I was just going over my first post and realized that it was a little too vague to generate interest, so I thought I'd give more details as to what I want to accomplish, so here's an outline of some of the features that I want to put into the program.

1.  When running, the program will have to be aware of the activities and applications running on it's host computer.  This means that it will be aware of the time, the date, whether the user is logged on, etc.  (Other programs have already accomplished this.)

2.  When parsing user input, the program will need the ability to understand the basic parts of speech of the English language.  I plan to accomplish this through user-editable text files or compressed text files.

3.  Text files or compressed text files will be the primary brain storage.  This will make the program fully configurable by the user.  Scripts will need to be written that allow the program to "understand" what the text files mean.

4.  Rather than trying to compile all of the data used by the program into a single file, the program will use many smaller files for pre-installed and earned data.  When the program is not running, the user will be able to add more knowledge to the program's files and the program will be able to utilize the data once it starts up.

5.  Language parsing will be handled by using text files for different parts of speach.  For instance, the adjective "pretty" would be listed and defined in the "adjectives.txt" file.

6.  The language parser will read user input and determine what form of input the user is stating.  For instance, if the AI sees a "?" at the end of a user's input, it will know that the user is asking a question.  Next, it would search a text file such as "QuestionQual.txt" to identify the first word of the question the user has asked.  Qualifiers in the "QuestionQual.txt" file would include "When, Who, How, Why, Where, What, Do, Can, May, Will, etc."

7.  This compartmentalized data structure is much like that used by the human brain.  For example, scientists are pretty much certain that the brain uses one area for sight and several areas for comprehending what is seen.  There is even a medical condition that causes people who have it to not be able to recognize faces.  I'd name the condition, but I don't have the issue of Wired Magazine that I read the article in on hand at the moment.

8.  The first step I would like to take is to develop the GUI, then begin programming the brain.  This would allow me to determine what graphical needs the program will have when it goes into alpha stage and we can begin testing it.

9.  The first graphical interface will use an animation system similiar to Microsoft Agent, but written to be SAPI 5 compliant, since the voices in SAPI 5 are much clearer and easier to understand.

10.  On a final note, in the text files where the word data is stored, the program will have pronunciation data, allowing it to pronounce the word according to it's actual pronunciation, even though the output that the user sees is written with the common english spelling.


This is by no means a complete list of features.  I'm constantly working on how the brain should function so that it works as seamlessly and as humanly as possible.  To be blunt, I want to have the team that helps me to develop this AI to have the honor of being the first group of developers to write an AI form that passes the Turing Test with flying colors.  It will be a long, complicated project, but I think it will be more than worthwhile.  Once the basic conversation engine and the GUI are produced, we can begin working on more functionality that will allow the program to perform actions, such as starting system software on a scheduled basis, notifying the user of upcoming appointments and special dates, and eventually being able to surf the internet for information to expand it's knowledge base, and finally, to comprehend what it says and knows.

But first, I just want to get the conversation engine completed.

Thanks,  I hope this clears up a little of the vagueness that was in the first post.

Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: admin on January 04, 2007, 05:49:47 pm
Hi Darkwolf and welcome to DG, hopefully you will get a better response now that the festive season is over, it's been quiet here lately.  One thing I wondered is have you looked at the UltraHal or Alice bots yet, as they can do some of these things already.
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 04, 2007, 06:00:32 pm
Actually, I've looked at UltraHal, ALICE, KARI, Verbots, and even Answerpad.  None of them can pull off a conversation convincingly enough, in my opinion.  They all have their high points, and honestly, I thing KARI is better than the others.  At least she normally stays on topic.

UltraHal was a little disappointing, and I've yet to find a graphical interface for ALICE.  I'd like to at least see one of those.

I grew up in the 70's though, and back then, they were promising us intelligent robots by the year 2000.  They're late!  So I thought, if you want something done, do it yourself. 

I actually have a business plan in the works for what to do after the open-source engine is developed.  Basically, once the open-source is finished I'd like to get the people who worked on it together, form a company, and then develop the open-source into a more robust engine for commercial purposes.  Once that's done, I want to move right into developing robotic bodies and teaching the AI to interact with it's surroundings.  I'm thinking probably 15 to 20 years long term, but eventuallly, I really would like to see the robotic maids we were all promised in the 70's.

I know it all sounds like one big pipe-dream, but the only country that even seems to be interested in developing this kind of technology seems to be Japan.  It's time for the rest of the world to catch up in the robotics field, and if no one else will lead it, then I'll grab the flag and start waving it until I get someone's attention.


Thanks for the support, and here's to a long running project!       :cool

Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: admin on January 04, 2007, 09:51:20 pm
Well I truthfully do not think it is a pipe dream as groups are getting a lot further in this field nowerdays, although there is still a lot of ground that is untapped, unchartered and unprogrammed.  See the 'Computers vs Common Sense Video' (http://www.digitalgirl.co.uk/dgforum/index.php?topic=1034.msg4407) thread to guage how advanced some have become already.

I would be interested in helping, but my own interest really lays with the graphical side which is later in your plans.  I can see my interest in this as being long term though and I will maintain some net presence either here or with similar ventures, so maybe our paths will cross.

Hopefully this site and the collection of resources will continue here or in some guise longterm too, so I would check back or repost ocassionally here as I do know some programmers pass by and may be interested in this kind of project.

Freddy

(I stickied this topic)
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 05, 2007, 07:12:10 pm
Thanks for the sticky!  I really appreciate it.  It would be rather nice if we could get everyone here involved with the project I have in mind.  Maybe we can call it the DigitalGirl AI Development Project or something.  Open source is definitely the way to go with AI, just as it was for the Linux operating system.  With open source, we can get AI to the masses!

Thanks!
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: FuzzieDice on January 11, 2007, 11:49:42 pm
Open source AI sounds like a nice idea! I haven't the time, unfortunatley, to even read your posts on this though. I just glanced at a few things.

Right now I'm busy with work projects (database-driven web applications) and haven't got the time to write code.

I *was* going to code my own AI. But I rather keep that one to myself becuase I'm getting rather skeptical of others knowing all my ideas.

I also offered to help code NICK, a new neuro-network AI system that Avi is creating. Though he is working on it himself at this point, recoding it all. I am sure he'll come up with something interesting and good. :) While I'm waiting on that release, I am just trying to keep up with work and all that.
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 12, 2007, 03:48:31 am
WOW!

I've been looking for the site for days!  I didn't see the message about the new page!  DOH!

Alot has been happening since I last posted.

I've gotten back into trying to learn the Python language.  I've done a small bit of programming in it before, so hopefully, I'll pick it up quickly.

With Python, I've begun programming my AI project.

For starters, the project has a name now.  Its called the ARIEL Project.  ARIEL stands for Artificial Reasoning Intelligent Electronic Lifeform.  To be honest, I named an artificially sentient computer Ariel in a story I wrote a long time ago, and wanted to use the same name for this new program.

I've already got the basics of a user interface GUI written in Python.  Now, I just have to get the AI written and get it to work with the interface.  I've attached a screenshot below.

I'll keep updating as things progress.



Darkwolf

Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: admin on January 12, 2007, 12:40:54 pm
Gald you found us :)  I notified most of the usual haunts, but was worried some of the newer members would lose us, so thanks for looking and sorry for the trouble but I was forced to make some changes.

Python is one of those things I want to get into more, I have only glanced over it when using Blender as that uses it quite a lot behind the scenes.  It's very popular and I don't recall seeing a bad word against it, so good luck.  Let me know if you want any help on the graphical side, but I like the robot image you used already.

What kind of avatar is it going to use - I'm just wondering if it's going to be possible to run the Haptek player using Python?

Just thinking...there will be more to do than just programming, things like compiling databases and some researching...so in a more general sense count me in.  If I can find some time I will look at Python, but I can definately help on things like Haptek if you use it (I've been writing Visual Basic apps for Haptek figures this past few months).


Freddy
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 12, 2007, 06:14:52 pm
Thanks Freddy!

I love the look of the new board, by the way.  It's got a nice, clean layout.  As far as interfaces go, I'm still playing around with ideas.  The one in the pic is basically just a still pic with some text boxes.  I've got to do some checking to see if there are any references for Haptek programming in Python.  Haptek would be the way to go, but I can't remember if they have licensing requirements for using the player in apps. 

As for Python, it's very easy to learn, once you get into it.  From everything I've read, it's even easier than Visual Basic, and tons easier than C or C++.  It's also considered to be a good replacement for LISP, which was built to be an AI programming language.  Hopefully, all will go well.

On the data side, I've found a Python wrapper program that allows Python to use AIML, so I can at least (I hope) learn to add ALICE functionality to ARIEL.  I've also found a wrapper program that gives Python access to WordNet, the Princeton Cognititve knowledge base.  If I can get all of that working, I might just be halfway to creating the AI I want to create.

The hard part is learning enough Python to figure out how to write the various modules and classes I'll need to put the main program together.  Honestly, I'd rather build everything in all at once so I don't have to go back and keep hacking to add more things in later, but I don't think that'll happen.

I'll keep everyone posted!

Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 12, 2007, 11:45:29 pm
Hello again...   :)

Things may have just gotten a whole lot more interesting.  I found an open source 3D engine like nothing I've seen so far.  The best part:  It was created by Disney, then released to Carnegie Mellon to be produced for open source.

To see more, go to http://www.panda3d.org/ - I've already downloaded the engine and I'm playing around with it now :) .

Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: admin on January 13, 2007, 01:30:17 pm
Great find, I am looking into it too :)  I really like the simple site layout too.
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: FuzzieDice on January 14, 2007, 01:29:51 am
I've used Python in Paint Shop Pro to code macros. I do like it, but you do have to be careful to indent your lines for the interpreter to work.

Python, Perl (which I also know), Ruby - all scripting languages. They are not OOP which makes them easier to learn than Visual Basic, C++, et. el. But they have the downside of being potentially slower (more on that in a moment) than compiled programs, if one runs the script via the interpretor. Python CAN compile the sources though. I think Perl can I'm not sure.

I've taken courses in Visual Basic.NET and have used Visual Basic 3, 5, and 6 (all of which I also have, though not installed). On Wednesday is the start of my C++ course. THAT is what I hope to use for an AI system. This is becuase I want to maybe do this in Linux. Python can be done in linux too, true. But I don't know maybe I'm old-skool, but I always thought C++ would have the libraries for database access. Does Python have database access libraries? I forgot if it does or not. I know Perl does.

I have heard of people trying to make AI systems in LISP and Prolog as well. I have no faimilarity of those. Pascal has gotten "old" I guess (anyone still use Pascal?) but that used to be popular at one time. I have briefly looked at pascal code at some times but haven't actually wrote any myself outside of Basic09 (an OS-9 Level II (for Tandy Color Computer 3) programming language that was a combination of BASIC and Pascal).

It USED to be that what language one used for AI programming determined if it would even have a chance or not. But now days, with the current versions of different languages, most are pretty powerful enough to do the trick, with some clever coding and the right libraries. I think Perl is also great because of it's natural uses of regular expressions and pattern matching.

Good luck with your project. I would be most interested in keeping up with what you are doing and how well it's going with Python. Are you using a database for storage? Which kind? (MS SQL, SQLLITE, MySQL, Postgres) Are you using Linux or Windows?

I think more could (and should) be done with Linux in regards to AI too.
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 14, 2007, 07:38:33 am
I've taken courses in Visual Basic.NET and have used Visual Basic 3, 5, and 6 (all of which I also have, though not installed). On Wednesday is the start of my C++ course. THAT is what I hope to use for an AI system. This is becuase I want to maybe do this in Linux. Python can be done in linux too, true. But I don't know maybe I'm old-skool, but I always thought C++ would have the libraries for database access. Does Python have database access libraries? I forgot if it does or not. I know Perl does.

Good luck with your project. I would be most interested in keeping up with what you are doing and how well it's going with Python. Are you using a database for storage? Which kind? (MS SQL, SQLLITE, MySQL, Postgres) Are you using Linux or Windows?



Hey Fuzzy!

Python does have database access libraries (and pretty much anything else you can think of - I know it has classes for MySQL.).  I chose python more because I've used it a bit in the past and it's a pretty easy language to learn.  Since it's also portable to all OS'es, I thought may once I get things up and running, I can find someone to make sure everything is portable to Linux and Mac's.  I'm on Windows XP (only because the games I play are available for it 100% of the time).

As far as database systems go, I haven't decided yet.  Right now, I'm still designing the way I want the program to flow, then I'll take it from there.  Python has a basic functions to create, read, and write text files to hold pretty much any type of data, and I'm hoping that will fit the bill for the ease of extending the AI's knowledge with plain text files (keeping my fingers crossed).

I'll keep everyone posted on my progress, and hopefully, I'll have at least a basic AI within a few months.

Cheers!
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: FuzzieDice on January 15, 2007, 11:27:37 pm
Text files are nice for sequential data and some portability. However, AIs will need to quickly search and retrieve, and also categorize data. For that you definitely will need to use a Database. I don't think a plain text file will cut it in that situation, if you want your AI to be able to "reason" and "learn" (ie. at least, even to emulate such behaviors).
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 16, 2007, 12:40:58 am
Good point FuzzieDice.  I've been thinking along the same lines myself.  I think the way I plan to disect the information may allow me to use some text files for small amounts of info, but for the large information bases, I most likely will have to go with databases.  I'll start looking at the databases available and see what will fit my needs.

Thanks!
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 31, 2007, 06:02:02 pm
Sorry for not posting in a long while.  I finally got my gaming machine back up and running, then went to do some upgrades and discovered I had a bad motherboard.  Three weeks and counting on getting everything finished on the new gaming machine.  Meanwhile, I've been re-exploring Oblivion, so haven't done much on the AI side.  I'm thinking about setting up team recruitment page on one of the open source sites to see if I can get some help.  It's taking forever to get things running on my own, as I'm still learning python and trying to put together a working gui as I go along.

I'll keep you posted on the developments.

Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on January 31, 2007, 10:32:56 pm
Okay, so I finally broke down and started programming on the project again.  Currently, ARIEL is in the text console phase...and a dos console at that.  However, she can speak using SAPI 5, and the voice, bot name, and user name are all configurable.  Right now, I'm reading up on another nice Python tool, the NLTK or Natural Language Tool Kit.  Basically, its an entire NL set already programmed and is compatible with Wordnet (which I plan to use).  Hopefully, this means ARIEL will at least start out with some serious vocabulary (if I can get the thing to work, that is).

I'll keep posting as progress continues.

Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on February 06, 2007, 03:20:57 pm
Hi gang!

Project ARIEL is unfortunately going to have to go in a different direction for the time being.  At the moment, I'm looking into implementing the project as an online chatterbot using the haptek interface.  Once I find somewhere to host the bot and get it coded (which should be much easier than the prior goals for ARIEL), I'll post a link here.

There are several reasons for this:

1)  I'm using the chatterbot to learn more about the current AI field.
2)  I'm using the chatterbot to learn the haptek interface.

And last but not least:

3)  After ALOT of reading, studying, and contemplating, I've come to the conclusion that what I want to accomplish will require more than software.  I've been reading up on this almost non-stop for several days (one tiny benefit of disability is free time).  The software that I would need to build would take more processing power than most desktop machines today can handle.  Unless I suddenly win the lottery, there's just no way I can afford to build the necessary hardware.  At this point, I'm thinking that the only way we're going to get past the chatterbot stage is with distributed hardware and some serious programming.  I know enough about distributed hardware to get myself into trouble, and enough programming to do some simple software.

Here's the reasoning behind my thoughts:

The human brain and consciousness are more than simple software.  The brain is a distributed network of hardware, with interfacing software.  Each individual piece of hardware is running a simple process, which it shares with other pieces of hardware in it's cluster.  Each cluster then collaborates with other clusters to form the "big picture" of understanding and reasoning.  The sheer number of processes running at one time will require a much larger processing capacity than what a current desktop computer can handle - and this would just be for true understanding of human language.  Add in other processes, such as image recognition and cognition and the processes grow exponentially.

Don't misunderstand me - I still think that a true AI is possible using today's technology.  For the time being, however, it's beyond my means, and most other people's means.  To truly get something like this off the ground would require large sums of money and years of collaborative work among programmers, neuro-scientists, and psychologists before we can get something approaching true Artificial Sentience.

The ARIEL Project is not dead!

I still plan to keep moving forward with the project, but at a much more realistic pace.  First, I need to start with fully grasping the current AI practice of mimicking human conversation and pattern-matching.  From there, I'll hopefully be moving into the areas of building small blocks of processes, one at a time.  Just as certain areas of the brain carry out single processes and then feed this information to other areas, each of these smaller blocks will communicate with other blocks of processes, until the whole begins to take shape.  Natural language processing will be the first goal of this second stage of development.

I believe at this point, that a single, multi-processor computer can serve as a single hardware block for each process of handling natural language.  I'm not talking about super-computers here.  I'm talking about relatively inexpensive systems that can serve as a single block or node in the overall network of hardware to process simple functions rapidly and efficiently, then pass the processed information forward to other hardware and software for higher-level processing.  A single, dual-core processor may be able to handle a single process.  On that side, I'll just have to wait and see how things develop.

Here's an example of what I'm talking about.

Let's say that the user enters the statement, "Dogs are mammals."  This input is parsed via a single hardware/software node to determine parts of speech.  The subject "dogs" is sent to a lower-level hardware/software node, where it is processed against a database.  If the information pertaining to dogs is not in the database, then the word is added and flagged for future processing.  The verb "are" is sent to another hardware/software node, where a similiar process takes place.  The noun "mammals" is sent to still another hardware/software node where it too is processed.  These results are then sent to a higher-level hardware/software node that flags "dogs" as belonging to the class "mammals".  The final process is that a final hardware/software node is told that the information is processed, and it in turn notifies the user that this has occurred.  The distributed system has just learned the fact that dogs are mammals.

Of course, this information can be added beforehand, so that the entire system already knows this fact before a user becomes involved.

Now, I'm sure some of you are going to say that this is an overly-complicated view of how the process might work.  You may be correct, but I'm looking at it from the point of view that human language is extremely complicated.  We still don't quite understand how we ourselves understand it.  A complicated system requires another complicated system to comprehend it.

My opinion is that by using this distributed network of hardware and software packages, it will allow the entire network to begin to learn and comprehend.  Artificial sentience won't take place on a single machine, but on a network of machines and the network will become the sentient entity.

As I said, the ARIEL Project is not dead.  Far from it.  It is only beyond my financial means and programming capabilities at the moment.

I come from a hardware background (many years in the field), although I am a bit behind on current hardware tech, since I was forced out of the field due to my illness.  By taking a step back and re-analyzing the problem, I can take the time to reaquaint myself with the latest hardware and sharpen my programming skills at the same time.  While doing this, I can also build a firmer paradigm of how I believe artificial sentience can be achieved.

In the long run, the ARIEL project will benefit from this shift in thinking.

As always, I'll keep everyone posted on how things are going.

Thanks,
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: admin on February 06, 2007, 08:35:44 pm
Interesting stuff.  I too do not believe we're going to see this kind of AI on the humble desktop for a long time yet either, but I could expect pc's to operate as nodes in a distributed system - something like the internet already is, but perhaps with an engine like the Cyc system turning the cogs of a technological intellegence.

The net already contains a large sum of human knowledge, so really with the knowledge in place, what is needed most is a method of extraction.  So, things like data mining come to mind and in preference to teaching from scratch, afterall if there already is a massive knowledge base, then why not use it.

I guess it all depends on the scale of what anyone wants to achieve.

I think this is also part of the process of miniturisation and streamlining - the continued reduction in the scale of technology but with the retention or increase in the level of information and the ease through which it is obtained.  Thirty years ago you would bet a computer scientist would swap his room-filling mainframe for an Apple Mac to do the same thing.  Information is becoming easier to obtain - you no longer have to invest a pile of cash into a massive library like you would if you were a Victorian, now it is at our finger tips.

I'm not sure if much has fundamentally changed, but the passion and belief that something fantastic can be created is alive and well.
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on February 06, 2007, 09:39:47 pm
Actually, some people already consider the internet to be alive.  ;D

Basically, they say that it has knowledge, can self-repair, and grows...must be alive.  I disagree, but I have heard the arguement for that.

BTW:

Here's the link I promised to the ARIEL Pandorabot:
http://demo.vhost.pandorabots.com/pandora/talk?botid=d38a302a0e3604e0 (http://demo.vhost.pandorabots.com/pandora/talk?botid=d38a302a0e3604e0)

If the haptek doesn't like to load, or the top banner fails, blame my free webhost...I hate using free hosts, but what can I say, I'm broke.

Later!
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on February 07, 2007, 06:12:37 pm
Okay, here's an update on the ARIEL Pandorabot version:

ARIEL is no longer using standard Alice AIML.  I'm in the process of rewriting the AIML to give her a unique personality.  Right now, she doesn't have alot of AIML in place, but she can chat to some degree.

I would consider it a great favor if some of the folks here would go and chat with her.  That way I can review the logs and see what needs to be added or changed.  Don't hold back.  Put her through her paces.  That's the best way I know how to build her knowledge base.

As usual, I'll keep posting updates as I go.

One last note.  Does anyone know of a free online chatbot system that actually has a learning AI?  I've looked everywhere and can't find one.  Even though the pandorabot system is educational, I find it very limiting.

Thanks!
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: admin on February 07, 2007, 07:17:53 pm
Taking shape :)  I tried chatting but didn't get far, good luck with coding her, and do keep us posted.  She seems to hang up sometimes and stops responding I found.
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on February 07, 2007, 09:43:57 pm
Hey folks.  Another update!

I think the hanging issue is either with the pandorbots server or with my web server.  I get the same issue and I also get it with the pandorabots site while working on her.

Okay.  Now for the update.

For the past few hours, I've been coding Ariel to respond to people seeking to have a sex chat with her.  I had no idea how many combinations of sex topics this would involve and I'm sure there are some I still haven't even thought of.

Here's your mission, should you choose to accept it:
Harass Ariel about sex.  I know that sounds weird, but its the best way I can think of to make her pervert-proof.

Try not to be shocked when she responds with some serious attitude and tells you off in a very unfriendly manner.

Don't worry, the logs are confidential and will only be used to program more information into Ariel.

I've also added in a modified Atomic.aiml file, which I'm still working on.  She should respond to normal conversation more now.

Thanks for all of the help and the ideas!

Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: PaleRider on February 07, 2007, 09:59:05 pm
 :)Hi
 
    all is she say's to me is (I have no answer to that), :( She must have got hit in the head ,When you moved her.
     I'll go back later and try to reason with her. Just Joke'n. I read your post hope it all comes together for you.
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on February 08, 2007, 02:20:43 am
Hi!

Right now she'll only respond to certain questions because I have all of the standard AIML turned off until I can go through it and see what I want to keep and what I want to throw out.

Ask her a little about herself (Where she's from, who her favorite band is, etc) and she should start making a little sense.  I've also changed the default "no answer" replies to give her a little more personality when she can't answer something.

I'm still in the process of basically rewriting alot of the AIML and creating quite alot of it from zero.  In a few more days her brain should start to be more responsive.

As usual, I'll keep everyone updated on the progress.  You should here her cuss someone out when they try to proposition her for sex.  She's learning some quite colorful language.  ;D

Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on February 13, 2007, 12:32:13 am
Hello again!

Just a quick update.  I've not really spent much time on the ALICE version of ARIEL in the past few days.  To be honest, I'm more than a little frustrated with the AIML framework.  I've probably edited aboout 1000+ lines, although not all of those are present at the moment.

What I really want is a learning AI, but I haven't found a web-based learning bot to use yet.  Most of the ones on the web seem to be coded by the people who own them and aren't available for use or sale.

As far as the ALICE program, I've learned that I'm not at all partial to canned and scripted responses as they make the bot repetitive and capable of only a limited conversation.  I'll admit that with enough work, an ALICE bot could possible seem to be intelligent, but I think it would take years of work in AIML before the bot could even come close to passing the Turing test.  The only way I can foresee an ALICE passing the test is with millions of AIML entries covering pretty much every semantic response possible in the human language.

I guess we're still a decade or more before a software implementation can even get close to passing the Turing and even further off before we'll have a truly intelligent bot.

<sigh>

Oh well, I'll keep looking and working until I get a bot that is capable of what I'm trying to accomplish.  I can see the years stretched out in front of me, but at least there is a goal to focus on.

Thanks!
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: admin on February 14, 2007, 12:28:40 pm
You may not be far off on the time scale; as I have heard 2020 suggested by a few ai makers as a date that the Turing test may be passed - it's anyone's guess I think though.
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: darkwolf77 on February 14, 2007, 07:30:31 pm
Hi again  :)

For some reason, I decided to work on the ARIEL ALICE version today and added some new content.   She can now talk a little about her favorite author and her boyfriend.  I also tweaked a few things here and there to make her a bit more interesting when she gives a default "I don't know" type of response.  I noticed in one log that a chatter asked her "who was talking to her" when she said that someone else was talking to her over there.

Basically, I'm just starting to put in topic changes and paths that keep her focused on what's being discussed.  The biggest challenge I have with building an AIML brain is making sure that she doesn't keep giving canned responses continously.  That means building a lot of random responses to every question or comment.  To put it mildly, it's a lot of tedious work.

Keep on building the AI dream. :)
Darkwolf
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: admin on February 15, 2007, 07:37:37 pm
Seems to be working okay now, it only hung once.  I can report her fav band is ZZ top :)
Title: Re: I'm looking for programmers to write code for a new generation desktop AI.
Post by: Maviarab on June 21, 2007, 10:21:47 pm
Darkwolf, hope your still about and that you subscribed to this thread so get notified of new replies.

What is currently happening with ARIEL?

Are you any closer to what you originally wanted?

Keep us updated  ;D