Personality

Hey, I've been working lately on a bot that can learn a language from scratch. The interesting bit is that this one actually knows what it's saying. The idea is that the bot learns to speak about geometrical shapes. It shows a shape on a window and you input some text describing the shape. OK, it is a very limited subject of conversation and possibly quite boring, but it's a first step towards having a computer that can actually speak instead of faking it.
Right now, my bot can learn to describe shapes in grammatically correct sentences. I've tried teaching it in English and Spanish, and both have gone down fairly well. Now, I'd like other people to test it (preferably people who can speak other languages). If anybody is interested, please send me an email to lu_si_na@yahoo.com

Thanks proff. Now everyone will be able to meet my weird little family! LOL

Weird and proud of it!

Doly, I've been trying to do that too. Anybody know another language, chat with me.

I just tried to chat with you..lol first you hung up on me for speaking French, then when I got you to come back, you said you didn't like French, and hung up on me every time I said anything about foreign language lol I'll try again later...

I am working on a new bot called Descartes Demon. He's at the Sophomore Stage, but eager to chat.

I'm sorry Ezzer and whoever else tried to chat with me...my language keyphrases didn't get triggered. They should now. Sorry about that. French away.

How do you get your foreign language past the spell checker and preprocessing?

Usually, you don't.

You can't, it will be seen by other bots as nonsense, so that can be fun...

I see. Educated babblebots.
We really need to be able to mark words and phrases so they won't be wrongly 'corrected' or altered by the preprocessing.

I agree, Joe. I posted in bug stomp about Julie's preprocessing incorrectly changing "your" to "you are" when my bot is set up to differentiate fine on her own, and I don't want spellcheck to correct it anyway....plus in this case it should have stayed "your". Maybe if we had a sort of 'custom dictionary' or 'do not correct list' as part of AI intitialization (I picture it working kinda like defaults)...hmmm