Hi! My name is Manuel and i'm new to this forum. English is not my native language so please excuse my many mistakes to come. I have signed on this forum because I have some concepts and theories on creating AI, that I want to present to you and hopefully something good will come out of it. My goal is to create an open source project that hopefully will deliver true AI in the years to come (hehe a lot of hopes here). So, in the next weeks and months I will present to you my theories, logic schemes, concepts, ideas, etc... Please join me on this endeavor. Here we go, off to the next level.

The first and biggest (the most important) concept of my theories : YOU CANNOT HAVE TRUE AI (OR AT LEAST TRULY FUNCTIONAL AI) WITHOUT MEMORY!
So when starting an AI programming you should start with a Memory System and a Mnemonic System. What i mean by a Memory System and a Mnemonic System - a way to memorize and store (also classify data) information that can be easily accessible. The system should be able store raw data and also (very important) along with the raw data, create and store Logic Links (or relations) between that data. For example - if i am to query the word "house", the system should come up with the relevant raw data but also with things related to it like "chair" or "watching tv" and even "clouds" (why clouds? - because they float above the house - that's how far the system should be able to create relations).
I know what you will say right now - Great, you just presented us with a chatterbot system that we already have! Don't worry this is just an outline, i will come with diagrams and more in-depth presentation.
What i wanted to convey to you is that if you create just a good enough Memorizing System (or Storing System - call it whatever you like) you could have a nice conversation with the computer, almost indistinguishable from a conversation with a human (given that we have enough data stored) - and we haven't even started creating AI yet.
I have tried the best chatterbots on the web and i find them really really really dumb, and the main reason for that is that they do not remember (or store and create logic links and relations from the data that i'm feeding it to them) almost any of the conversation that i have with them.
That's it for today. Next time i will give you a more in-depth presentation of some of the mechanism behind the Memorizing System.
I'm waiting for your feedback. Thank you!

Is this related to AI for games? If not, you are likely in the wrong place.

Hi Manuel,

I'm sure the crew over at Chatbots.org AI Zone forum would like to hear more about this at http://www.chatbots.org/ai_zone/ ..

Cheers,

- Marcus Endicott
http://twitter.com/mendicott

All I have to say is... what an honor to have a distinguished mind such as yourself in our presence. I think we should all remember where we were on June 21 at 1:50 AM, because I believe your post marks the beginning of an era. That you can create an AI far superior than all the dumb AIs currently existence is obvious -- that you haven't yet is of no consequence -- and we all are waiting in rapt attention for you to unleash your superior AI into the world, thereby ushering in the Age of the Jackangel.

A sincere welcome from your humble admirers.

Is this related to AI for games? If not, you are likely in the wrong place.

Isn't supposed AI to be somewhat universal? I does apply to games, virtual pals or whatever...

Hi Manuel,

I'm sure the crew over at Chatbots.org AI Zone forum would like to hear more about this at http://www.chatbots.org/ai_zone/ ..

Cheers,

- Marcus Endicott
http://twitter.com/mendicott

Thanks for your reply. But i'm not trying to create just a chatterbot. I was saying that just by implementing a good memory system you get a very good chatterbot.

Isn't supposed AI to be somewhat universal? I does apply to games, virtual pals or whatever...

Nope. There are techniques that robotics uses that games doesn't and vice-versa. Techniques that Watson uses that games don't. Plenty of division points.

Thanks for your reply. But i'm not trying to create just a chatterbot. I was saying that just by implementing a good memory system you get a very good chatterbot.

No you don't. Memory systems can't parse the vagaries of human language, understand relative context, etc.

I'm assuming you have read Expressive Processing by Noah Wardrip-Fruin and The Most Human Human by Brian Christian? If you had, you would know that it takes more than a memory or knowledge representation layer. If you haven't, then you are starting from scratch anyway.

All I have to say is... what an honor to have a distinguished mind such as yourself in our presence. I think we should all remember where we were on June 21 at 1:50 AM, because I believe your post marks the beginning of an era. That you can create an AI far superior than all the dumb AIs currently existence is obvious -- that you haven't yet is of no consequence -- and we all are waiting in rapt attention for you to unleash your superior AI into the world, thereby ushering in the Age of the Jackangel.

A sincere welcome from your humble admirers.

Always love a snide remark, and I appreciate you taking the time to write that. I would love to not have admirers but people to work with on this project. I'm not trying to be a smart ass, i'm not saying that i have an improved AI system - i just want to present to you with another approach to AI. Time will tell if my approach is better or not. Frankly i find a little disappointing that in the year 2011 we do not have a true AI system. It's been like 15 years since we have enough CPU power and enough disk space to support AI and no real AI in sight. And I'm not going into conspiracy theories that say that there is true AI created but some governments banned it and blah blah blah...if that is true, then too bad - we have to reinvent it again...
I wish you all a good day! See you next time!

[quote name='Jackangel' timestamp='1308688527' post='4826130']
Thanks for your reply. But i'm not trying to create just a chatterbot. I was saying that just by implementing a good memory system you get a very good chatterbot.

No you don't. Memory systems can't parse the vagaries of human language, understand relative context, etc.

I'm assuming you have read Expressive Processing by Noah Wardrip-Fruin and The Most Human Human by Brian Christian? If you had, you would know that it takes more than a memory or knowledge representation layer. If you haven't, then you are starting from scratch anyway.



[/quote]

That is exactly the problem that i'm trying to address. That is what I'm trying to demonstrate - if you give me the opportunity. It is all about "Loose Logic" (what i like to call it) or Fuzzy Logic. We are not that complicated, our thinking and logic it is not that evolved, as a matter of fact we are most of the time illogical (yes i know, i sound like Spock) - and that it's exactly what an AI system needs - a Logic with some leverage, a Loose Logic.