LISA, the "Lexicon Integration via Software AI" Unit...an automated conversational software.  OK, so what does that mean?  Well, LISA is capable of not only responding, but learning...and that's the real purpose of LISA - to learn.  Any kid can make a program which can ask for an input "Hi how are you", and respond with an output "Good", with only a couple lines of code...and I mean, that's honestly all most chatbot programs really are - a series of If...Then statements.  And that's not artificial intelligence.  But does that mean that chatbots are useless...

No, quite the opposite really.  Chatbot technology is the foundation of AI, the backbone and framework for which the rest of the AI can build from.  Creating AI is no easy task, well, save for the few government secret projects, is non-existent.  But it is my belief that this is so because many projects start out attempting to dive straight into AI, from ground up, and I honestly don't believe that's an effecient way of doing things.

I tend to liken chatbot tech to a computer monitor.  Unless you have perfect typing skills with a history of zero typos, a perfectly outlined design document that states how to write every single line of code, and some serious dumb luck, you won't be able to code anything meaningful without a computer monitor.  If you make even the slightest mistake, you'll get an error and you'll have no way to know how to fix it without being able to actually see the code.  That's what chatbot tech is, the monitor, the visual aid, allowing you to physically see and hear how the AI feels and responds.

And that's where LISA comes in.  LISA is the speech/conversation module to a greater AI project yet to be developed.  LISA provides the user with feedback, eventually well thought out feedback, with meaning and emotion.  Having a response ready for a particular input is not artificial intelligence, but if LISA can understand the input - understand the underlying meaning of whatever it is the user is trying to convey - then LISA can design her own response.

The end of cause-effect (input-response) programming.

Before an AI unit can intelligently speak, that unit must first speak...period.  Once the platform has been set, however, you can slowly start introducing different elements.  LISA's first goal is to converse with the input-response scheme.  Next, LISA will learn to replace certain words with words of similar meaning...a thesaurus module.  At this point, LISA will also learn how to understand words by their meaning.  

Each word will have a "weight" associated with them; a good word, bad word, verb, noun, powerful, emotional, etc.  LISA will also track how often some words are used by the user, and apply a percentage "weight" value.  The "heavier" a word, the more LISA will tend to use it.  At this point, certain response structures can be built, where certain words can be replaced or added in, but the overall "weight" of the response must be about the same.

And I mean, that's it...the basic idea for how LISA will be functioning and why.  So take a look around the site, especially the "What's New" page to see current updates on the actual LISA...