Descargar Bot Libre de Escritorio, su propio bot para su PC o Mac
Self, AIML, and scripting : Evaluation flow, declaring new :response sentences and miscelany

RE: Evaluation flow, declaring new :response sentences and miscelany

por aazxcqwe publicado Jan 23 2015, 13:07

Thank you, I believe I tried assign, but I will try that again. As I said I set the 'if(#false,#true)' purposefully to return #false, because I wasn't able to maintain variable definitions. This should help.

Also, I forgot to mention that >5k bots is impressive!

Here is a plain English description of what I envision. I think I've tried to explain this, but I really appreciate the input and here goes again:

Big picture: I'd like the bot to learn from sentences, expanding on the NounVerbAdjective script. Possibly simplifying it, or at least learning every step of it by trying to extend it. Also, possibly delegating this out to an external API if the development is way past what I can efficiently code. Ultimately, I would like to be able to define macros or scripts by voice using the bot (a goal also in the BotLibre! descriptions). Defining objects and attributes is a first step toward this. Later the objects may become Android intents and attributes their parameters (or PC scripts). I realize this is a big goal, but I'm testing the waters trying to create the fundamental features (beyond the Self language) to allow this, in the most general and flexible way.

This example: I would like to be able to generally define a sentence's subject for more than defined (NounVerbNoun structure matches) in the way I see online sentence parsing tools do. I'm not sure if they have individual routines for each structure, or how many structures are necessary (there may only be a few, but then again language especially English can be more complex than it appears), which is why I contemplate using an external service similar to how the bots search Wiktionary and Freebase. The other approach, likely more powerful is to develop simple language scripts that enable the bot to extend its own scripts and sentence comprehension, similar to very early language learning in children.

This specific script: I'm trying to have a routine that checks if an association between a word (eventually identified as the subject) and all other words of the sentence. If no association exists, I would like it to proceed to create one through an "#is" relationship or else an "#is not" relationship if "is not" is included within the sentence.

After that is probably gets more complicated with different sentence fragments, but if there was a general method for the bot learning new sentence parsing procedures, I suppose that would be ideal. Maybe this is what the Comprehension or Learning Mode options do.
Even more generally, I think the bot needs to be checking if it understands how each part of a sentence relates to each other logically, and if not ask the appropriate questions to make the sentence 'make sense'. I would be surprised if there weren't some researchers out there with a method for this already defined, but then again maybe the most active development is being done in the private sector, or maybe we'll see it here first and someone can try for a Loebner!
Thanks again!

Id: 716509
Publicado: Jan 23 2015, 13:07
Actualizado: Jan 23 2015, 13:31
Respuestas: 0
Vistas: 1917, hoy: 1, la semana: 1, mes: 13
0 0 0.0/5